Add files using upload-large-folder tool

TASK_METHOD_20_GAP_AUDIT.md

@@ -1,6 +1,6 @@
 # Task Method 20-Result Gap Audit
 
-Generated: `2026-06-18T12:07:14+00:00`
+Generated: `2026-06-18T12:52:47+00:00`
 
 This audit is the explicit gap ledger for the 9-method x 20-task result matrix.
 It keeps missing cells visible while preserving the rule that a numeric score
@@ -9,8 +9,8 @@ requires a real task target and source artifact.
 ## Score Summary
 
 - Method-task records: `180`
-- Numeric scored records: `127`
-- Scoreless records: `53`
+- Numeric scored records: `143`
+- Scoreless records: `37`
 - Proxy-scored records: `4`
 - Source matrix: [`docs/data/task_method_20_result_matrix.json`](docs/data/task_method_20_result_matrix.json)
 
@@ -20,8 +20,8 @@ requires a real task target and source artifact.
 | --- | --- | --- | --- | --- | --- |
 | Minimal | minimal | 20/20 | 0 | 0 | scored: 20 |
 | Neural MLP | neural_mlp | 20/20 | 0 | 0 | scored: 20 |
-| 128ep Metadata Simple | metadata128_simple | 13/20 | 7 | 0 | scored: 13, unsupported_without_required_target: 7 |
-| 128ep Metadata NN | metadata128_neural_mlp | 7/20 | 13 | 0 | not_supported_by_metadata_only_package: 7, scored: 7, unsupported_without_required_target: 6 |
+| 128ep Aligned Simple | metadata128_simple | 18/20 | 2 | 0 | scored: 18, unsupported_without_required_target: 2 |
+| 128ep Aligned NN | metadata128_neural_mlp | 18/20 | 2 | 0 | not_supported_by_metadata_only_package: 2, scored: 18 |
 | 128ep Raw Simple | raw128_simple | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
 | 128ep Raw NN | raw128_neural_mlp | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
 | Qwen3-Omni v6 LoRA | qwen3_omni_v6_lora | 15/20 | 5 | 0 | not_evaluated_in_verified_package: 5, scored: 15 |
@@ -33,61 +33,45 @@ requires a real task target and source artifact.
 | Status | Count | Next step |
 | --- | --- | --- |
 | not_evaluated_in_verified_package | 33 | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| not_supported_by_metadata_only_package | 7 | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
-| unsupported_without_required_target | 13 | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
+| not_supported_by_metadata_only_package | 2 | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
+| unsupported_without_required_target | 2 | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 
 ## Scoreless Records
 
 | Task | Task label | Method | Status | Required evidence |
 | --- | --- | --- | --- | --- |
-| 01 | Action Recognition | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 02 | Procedure Step Recognition | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 | 02 | Procedure Step Recognition | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 04 | Next-Action Prediction | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 05 | Hand Trajectory Forecasting | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 05 | Hand Trajectory Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 05 | Hand Trajectory Forecasting | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 05 | Hand Trajectory Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 05 | Hand Trajectory Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 07 | Object Relevance Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 08 | Language Grounding | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 08 | Language Grounding | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 09 | Cross-Modal Retrieval | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 09 | Cross-Modal Retrieval | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 09 | Cross-Modal Retrieval | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 10 | Cross-Modal Reconstruction | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 10 | Cross-Modal Reconstruction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 10 | Cross-Modal Reconstruction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 10 | Cross-Modal Reconstruction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 10 | Cross-Modal Reconstruction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 11 | Temporal Order Verification | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 11 | Temporal Order Verification | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 12 | Multimodal Synchronization Detection | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 12 | Multimodal Synchronization Detection | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 12 | Multimodal Synchronization Detection | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 12 | Multimodal Synchronization Detection | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 13 | Long-Horizon Next-Action Forecasting | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 14 | Long-Horizon Next-Subtask Forecasting | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 15 | Interaction Text Prediction | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 15 | Interaction Text Prediction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
+| 15 | Interaction Text Prediction | 128ep Aligned Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
+| 15 | Interaction Text Prediction | 128ep Aligned NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 15 | Interaction Text Prediction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 15 | Interaction Text Prediction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 15 | Interaction Text Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 16 | Action-Object Relation Prediction | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 | 16 | Action-Object Relation Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 17 | Future Object-Set Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 17 | Future Object-Set Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 18 | IMU-to-Hand Pose Reconstruction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
-| 19 | Camera-View Synchronization Retrieval | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
-| 19 | Camera-View Synchronization Retrieval | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
+| 19 | Camera-View Synchronization Retrieval | 128ep Aligned Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
+| 19 | Camera-View Synchronization Retrieval | 128ep Aligned NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
 | 19 | Camera-View Synchronization Retrieval | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 19 | Camera-View Synchronization Retrieval | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 | 19 | Camera-View Synchronization Retrieval | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |

TASK_METHOD_20_RESULT_MATRIX.md

@@ -8,8 +8,8 @@ Legend: `score` = numeric task score, `proxy` = documented raw128 compact proxy
 | --- | ---: | ---: | ---: | ---: | --- |
 | Minimal | 20 | 20 | 0 | 0 | scored 20 |
 | Neural MLP | 20 | 20 | 0 | 0 | scored 20 |
-| 128ep Metadata Simple | 20 | 13 | 0 | 7 | scored 13, unsupported 7 |
-| 128ep Metadata NN | 20 | 13 | 0 | 7 | not supported 7, scored 13 |
+| 128ep Aligned Simple | 20 | 18 | 0 | 2 | scored 18, unsupported 2 |
+| 128ep Aligned NN | 20 | 18 | 0 | 2 | not supported 2, scored 18 |
 | 128ep Raw Simple | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
 | 128ep Raw NN | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
 | Qwen3-Omni v6 LoRA | 20 | 15 | 0 | 5 | not evaluated 5, scored 15 |
@@ -22,20 +22,20 @@ Legend: `score` = numeric task score, `proxy` = documented raw128 compact proxy
 | 02 | Procedure Step Recognition | score | score | score | score | score | score | score | score | not evaluated |
 | 03 | Action Boundary Detection | score | score | score | score | score | score | score | score | score |
 | 04 | Next-Action Prediction | score | score | score | score | score | score | score | score | score |
-| 05 | Hand Trajectory Forecasting | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
+| 05 | Hand Trajectory Forecasting | score | score | score | score | score | score | not evaluated | not evaluated | not evaluated |
 | 06 | Contact State Prediction | score | score | score | score | score | score | score | score | score |
 | 07 | Object Relevance Prediction | score | score | score | score | score | score | score | score | not evaluated |
 | 08 | Language Grounding | score | score | score | score | score | score | score | not evaluated | not evaluated |
-| 09 | Cross-Modal Retrieval | score | score | unsupported | not supported | score | score | score | not evaluated | score |
-| 10 | Cross-Modal Reconstruction | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
+| 09 | Cross-Modal Retrieval | score | score | score | score | score | score | score | not evaluated | score |
+| 10 | Cross-Modal Reconstruction | score | score | score | score | score | score | not evaluated | not evaluated | not evaluated |
 | 11 | Temporal Order Verification | score | score | score | score | score | score | score | not evaluated | not evaluated |
-| 12 | Multimodal Synchronization Detection | score | score | unsupported | not supported | score | score | score | not evaluated | not evaluated |
+| 12 | Multimodal Synchronization Detection | score | score | score | score | score | score | score | not evaluated | not evaluated |
 | 13 | Long-Horizon Next-Action Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
 | 14 | Long-Horizon Next-Subtask Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
 | 15 | Interaction Text Prediction | score | score | unsupported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
 | 16 | Action-Object Relation Prediction | score | score | score | score | score | score | score | score | not evaluated |
 | 17 | Future Object-Set Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
-| 18 | IMU-to-Hand Pose Reconstruction | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
+| 18 | IMU-to-Hand Pose Reconstruction | score | score | score | score | score | score | not evaluated | not evaluated | not evaluated |
 | 19 | Camera-View Synchronization Retrieval | score | score | unsupported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
 | 20 | Time-to-Next-Transition Regression | score | score | score | score | score | score | score | not evaluated | not evaluated |
 

data/artifact_index.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:52:48+00:00",
   "status": "pass",
   "artifact_count": 213,
   "missing": [],
@@ -290,8 +290,8 @@
       "surface": "repo_hf",
       "shows": "Runs simple metadata and neural MLP baselines on the same selected 96/16/16 episode split used by the Qwen3-Omni diagnostic pilot.",
       "exists": true,
-      "bytes": 73236,
-      "sha256": "76acae0de25d51413e7e6f11021163e7d9909cfe95d65bf6b02e74043d429e2d"
+      "bytes": 74368,
+      "sha256": "6f54bfb963d5102ebd61eb8f8b6d8f6919db673378c9d5940d89ec5ea6f3d4b2"
     },
     {
       "id": "task_suite_enhancement_128",
@@ -599,7 +599,7 @@
       "shows": "Machine-readable source-alignment pass/fail check for repo, website, and HF surfaces.",
       "exists": true,
       "bytes": 4432,
-      "sha256": "ae089cc0df132b63365e03b2157a488b5d1569567c0374d7621bcd347da62c9e"
+      "sha256": "8ddadfe15ba8779e82879f965ff50bceb9c573bc942c3ecf176fbf20e5faeaea"
     },
     {
       "id": "source_alignment_validator",
@@ -719,8 +719,8 @@
       "surface": "website_hf",
       "shows": "Stores normalized 20-axis radar values, raw task metrics, Qwen3/Cosmos overlay mappings, branch-card caveats, and explicit scoreless status records.",
       "exists": true,
-      "bytes": 230297,
-      "sha256": "437874b1633e73165e3300f55580394663a44759c848288e696859b98f8aad32"
+      "bytes": 229299,
+      "sha256": "30f338139df391c36941da0b759cc237366ee43d006bfff2d2e43481cc2d2a63"
     },
     {
       "id": "single_episode_task_model_radar_json",
@@ -730,8 +730,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable split radar for the one-episode Minimal and Neural MLP baselines, both scored on all 20 task contracts.",
       "exists": true,
-      "bytes": 50973,
-      "sha256": "38cb43512f2ac40feeb62333bdea89b3a55e5b48468beb8982cf22536f794ecf"
+      "bytes": 51064,
+      "sha256": "52001c8ac081b14827a8a55cae21da8fd32516f81365d7dda1047ef68096eef8"
     },
     {
       "id": "episode128_task_model_radar_json",
@@ -741,8 +741,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable split radar for selected 128-episode metadata/raw baselines and verified Qwen3/Cosmos branches, preserving explicit scoreless cells.",
       "exists": true,
-      "bytes": 186443,
-      "sha256": "55e758e8703f406889022976d0ba055181212305c9a7246e899463e0c3c3b554"
+      "bytes": 185447,
+      "sha256": "e9994f42a1e086411748e1233761c84a8dcd564898c216454a8872c2f4d4f213"
     },
     {
       "id": "task_method_20_result_matrix_json",
@@ -752,8 +752,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable 9-method by 20-task matrix where every method has 20 records and scoreless cells carry unsupported/not-evaluated reasons.",
       "exists": true,
-      "bytes": 129242,
-      "sha256": "64fb700d51f536edf11291799b6173cf9ae8dd7a41178aac348b8207ed4b1e42"
+      "bytes": 128794,
+      "sha256": "1bce6001518b314fc8a5e86eab56521aa9718d09d787765d10caee4d791e9809"
     },
     {
       "id": "task_method_20_result_matrix",
@@ -763,8 +763,8 @@
       "surface": "repo_hf",
       "shows": "Reader-facing table that separates 20 records per method from numeric scored axes, documented raw128 proxy scores, unsupported metadata targets, and model targets not evaluated in verified packages.",
       "exists": true,
-      "bytes": 4026,
-      "sha256": "55e949fc30419a52f7f5ec4dd9544a11b253b076f8e3637ec3e92b3d61a89aab"
+      "bytes": 3954,
+      "sha256": "01b21d83954f700e4b061e96b1f58c6af474d79a2caaff1bfcff4854b66722ca"
     },
     {
       "id": "task_method_20_gap_audit_json",
@@ -774,8 +774,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable 180-record gap ledger with numeric scores, scoreless cells, explicit status reasons, and next evidence needed before new scores can be published.",
       "exists": true,
-      "bytes": 46902,
-      "sha256": "2b64dbd013625852679f9b91d25c48d1ed197fec727883b4fe37088b2d594784"
+      "bytes": 35883,
+      "sha256": "9336756d67d2488a28c4bb9c282f65230031eeb8dddd087a11fd441d8e61539b"
     },
     {
       "id": "task_method_20_gap_audit",
@@ -785,8 +785,8 @@
       "surface": "repo_hf",
       "shows": "Reader-facing ledger that lists every scoreless method-task cell and the concrete target or model-output evidence required before it can become numeric.",
       "exists": true,
-      "bytes": 13387,
-      "sha256": "d33461eb704f8e92545b6b54d9fc509e617fbacc9ca9894ac851ca9c3dec0fec"
+      "bytes": 10286,
+      "sha256": "45969b72e9a3ff8c40d958ea819e725fd4df5d90424ccdffd1c64fd1a5152063"
     },
     {
       "id": "unified_task_model_radar_chart",
@@ -796,8 +796,8 @@
       "surface": "website_hf",
       "shows": "Compares minimal and neural MLP baselines across all 20 tasks, with Qwen3/Cosmos task-aligned model overlays.",
       "exists": true,
-      "bytes": 51953,
-      "sha256": "19c001f10319946ef0e4921064f8a012836f29e7c8b272f900c257169faf46a1"
+      "bytes": 53553,
+      "sha256": "ec9a8bf0f5814106ddb8e62d0941c7cc07d1b8a29323a61a400319ffe6bd3485"
     },
     {
       "id": "single_episode_task_model_radar_chart",
@@ -818,8 +818,8 @@
       "surface": "website_hf",
       "shows": "Separates the selected 128-episode methods: raw-feature simple/NN as complete 20/20 scored polygons and metadata/Qwen/Cosmos as task-aligned overlays.",
       "exists": true,
-      "bytes": 45937,
-      "sha256": "b504b1b9c5cad0caa8c822d5bb2971c1b708251cf7b9ef587a92db2c12751e97"
+      "bytes": 47540,
+      "sha256": "0c2283a04fe401851b8b313de3ba383d24185262f4c6500d12fa0a3b8c0c4443"
     },
     {
       "id": "unified_task_model_radar_builder",
@@ -829,8 +829,8 @@
       "surface": "repo_hf",
       "shows": "Regenerates the direction-aware radar chart and machine-readable metric overlay JSON.",
       "exists": true,
-      "bytes": 52388,
-      "sha256": "f4803360cfd02383a1942a93a5845308db936b479a5b906719e46e192f3ef142"
+      "bytes": 52743,
+      "sha256": "e081f88e9f31934b24820c5cbffb957bb235a3275f553e573ab44e5c3d03c99a"
     },
     {
       "id": "task_method_20_gap_audit_builder",
@@ -906,8 +906,8 @@
       "surface": "repo_hf",
       "shows": "Rerun of JSONL metadata/text simple and neural baselines over the selected 128-episode multiscale dataset; supports radar overlays on JSONL-supported task axes.",
       "exists": true,
-      "bytes": 109248,
-      "sha256": "5e7f3085be5012eb3dda46f9c7b5b7c0ae22d6a0fbce71d6e99dd317fecc12af"
+      "bytes": 124232,
+      "sha256": "dba221a6ed8a6a84602dc21a1055cbb4444c03775f74b55e5d72861941820ac8"
     },
     {
       "id": "a100_128_raw20_task_baselines",
@@ -1310,7 +1310,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 994053,
+      "bytes": 1059014,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -1620,7 +1620,7 @@
       "shows": "Reader-facing comparison of the single-episode task suite, 128-episode aligned baselines, Qwen3-Omni packages, and Cosmos3 future-window branch.",
       "exists": true,
       "bytes": 15999,
-      "sha256": "30053bdea6c417ab02f98d99d8e80cd7e304bc3a9dfacbf599139d3221c02c8f"
+      "sha256": "dd65ae9077acbce91870b182d701db367a9c79eb287aeee2a1e165ec4915e5f3"
     },
     {
       "id": "omni_model_comparison_json",
@@ -1631,7 +1631,7 @@
       "shows": "Machine-readable comparison of the current result versions, per-task aligned baselines, verified Qwen3 packages, and Cosmos3 package.",
       "exists": true,
       "bytes": 81866,
-      "sha256": "1c9d4ba370661b0e0cb7104e9a51abdc3fe91a440ae86e748b10b719d1d613cc"
+      "sha256": "dd7a599117defcc1fd783c3134b6b3fc92f2ec2190ea517624cb215b931bd87a"
     },
     {
       "id": "cosmos3_nano_verified_summary",

data/episode128_task_model_radar.json

@@ -1,19 +1,19 @@
 {
   "title": "128-Episode 20-Task Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "description": "Selected 128-episode metadata/raw baselines plus verified Qwen3/Cosmos branches. Every method has 20 records; numeric scores appear only where the public artifact produced that task target.",
   "task_count": 20,
   "method_count": 7,
   "method_task_record_count": 140,
-  "scored_method_task_count": 93,
+  "scored_method_task_count": 103,
   "normalization_policy": {
     "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
     "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "source_unified_radar": "docs/data/unified_task_model_radar.json",
@@ -21,50 +21,50 @@
   "series": [
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -205,7 +205,7 @@
           "raw": 0.008252821966746326,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008252821966746326,
@@ -216,7 +216,7 @@
           "raw": 0.004175793689174209,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004175793689174209,
@@ -296,7 +296,7 @@
           "raw": 0.00019512195121951218,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.00019512195121951218,
@@ -307,7 +307,7 @@
           "raw": 7.207207207207208e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 7.207207207207208e-05,
@@ -387,7 +387,7 @@
           "raw": 0.29652162550029315,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.29652162550029315,
@@ -398,7 +398,7 @@
           "raw": 0.4841733292368365,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4841733292368365,
@@ -478,7 +478,7 @@
           "raw": 0.006514774539765508,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.006514774539765508,
@@ -489,7 +489,7 @@
           "raw": 0.004910507980164745,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004910507980164745,
@@ -566,26 +566,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 8.817333221435547,
           "metric_key": "mpjpe",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.012231610603598841,
+          "raw_text": "8.817",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.429434210062027,
           "metric_key": "mpjpe",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.25114484128127007,
+          "raw_text": "0.4294",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.2729249894618988,
@@ -660,7 +660,7 @@
           "raw": 0.4381481308057444,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4381481308057444,
@@ -671,7 +671,7 @@
           "raw": 0.5682695682695682,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.5682695682695682,
@@ -751,7 +751,7 @@
           "raw": 0.17764578833693304,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17764578833693304,
@@ -762,7 +762,7 @@
           "raw": 0.18662723837686876,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.18662723837686876,
@@ -842,7 +842,7 @@
           "raw": 0.002332374220713973,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.002332374220713973,
@@ -853,7 +853,7 @@
           "raw": 0.008236799389123917,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008236799389123917,
@@ -930,26 +930,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.002587692579254508,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.002587692579254508,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.0026067993603646755,
           "metric_key": "mrr",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0026067993603646755,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.003459817497059703,
@@ -1021,26 +1021,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": -190.66106203944798,
           "metric_key": "r2",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-190.66",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": -0.43481132003942147,
           "metric_key": "r2",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-0.4348",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": -1.3450960391924882,
@@ -1115,7 +1115,7 @@
           "raw": 0.4198864140782312,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4198864140782312,
@@ -1126,7 +1126,7 @@
           "raw": 0.8252408266656923,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.8252408266656923,
@@ -1203,26 +1203,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.49980060227663614,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.49980060227663614,
+          "raw_text": "0.4998",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.7773773780941162,
           "metric_key": "f1",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.7773773780941162,
+          "raw_text": "0.7774",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.4958867673901769,
@@ -1297,7 +1297,7 @@
           "raw": 0.004579592783699693,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004579592783699693,
@@ -1308,7 +1308,7 @@
           "raw": 0.0029821307969142615,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0029821307969142615,
@@ -1388,7 +1388,7 @@
           "raw": 0.0001206030150753769,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0001206030150753769,
@@ -1399,7 +1399,7 @@
           "raw": 2.086049543676662e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 2.086049543676662e-05,
@@ -1479,7 +1479,7 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
           "normalized_score": null,
@@ -1490,9 +1490,9 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1570,7 +1570,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1581,7 +1581,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1661,7 +1661,7 @@
           "raw": 0.17656983343047333,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17656983343047333,
@@ -1672,7 +1672,7 @@
           "raw": 0.17418550827844048,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17418550827844048,
@@ -1749,26 +1749,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.2294670194387436,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.18324815505876868,
+          "raw_text": "0.2295",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.2555866539478302,
           "metric_key": "mae",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.16452114110609004,
+          "raw_text": "0.2556",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.22941437363624573,
@@ -1843,7 +1843,7 @@
           "raw": null,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
           "normalized_score": null,
@@ -1854,9 +1854,9 @@
           "raw": null,
           "metric_key": "mrr",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1934,7 +1934,7 @@
           "raw": 624.8108520507812,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.016864874132806403,
@@ -1945,7 +1945,7 @@
           "raw": 41.4664421081543,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.25411768748242325,
@@ -2016,7 +2016,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2026,7 +2026,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2034,7 +2034,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2044,7 +2044,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2142,7 +2142,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2152,7 +2152,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2160,7 +2160,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2170,7 +2170,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2268,7 +2268,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2278,7 +2278,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2286,7 +2286,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2296,7 +2296,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2394,7 +2394,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2404,7 +2404,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2412,7 +2412,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2422,7 +2422,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2520,36 +2520,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -2646,7 +2646,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2656,7 +2656,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2664,7 +2664,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2674,7 +2674,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2772,7 +2772,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2782,7 +2782,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2790,7 +2790,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2800,7 +2800,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2898,7 +2898,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2908,7 +2908,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2916,7 +2916,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2926,7 +2926,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3024,36 +3024,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -3150,36 +3150,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -3276,7 +3276,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3286,7 +3286,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3294,7 +3294,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3304,7 +3304,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3402,36 +3402,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -3528,7 +3528,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3538,7 +3538,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3546,7 +3546,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3556,7 +3556,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3654,7 +3654,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3664,7 +3664,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3672,7 +3672,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3682,7 +3682,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3780,7 +3780,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -3790,7 +3790,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -3798,7 +3798,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -3808,8 +3808,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -3906,7 +3906,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3916,7 +3916,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3924,7 +3924,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3934,7 +3934,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4032,7 +4032,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4042,7 +4042,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4050,7 +4050,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4060,7 +4060,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4158,36 +4158,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -4284,7 +4284,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -4294,7 +4294,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -4302,7 +4302,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -4312,8 +4312,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -4410,7 +4410,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4420,7 +4420,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4428,7 +4428,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4438,7 +4438,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

data/omni_model_comparison.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Current Result Versions and Model Groups",
-  "generated_at_utc": "2026-06-13T18:14:42+00:00",
+  "generated_at_utc": "2026-06-18T12:52:47+00:00",
   "status": "pass",
   "version_count": 3,
   "model_group_count": 5,

data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -18,7 +18,7 @@
         "website_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:41:43+00:00"
+          "generated_at_utc": "2026-06-18T12:09:46+00:00"
         },
         "rendered_site_check": {
           "exists": true,
@@ -28,27 +28,27 @@
         "task_surface_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:25+00:00"
         },
         "source_alignment": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:45+00:00"
         },
         "scale_up_status": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:06+00:00"
+          "generated_at_utc": "2026-06-18T12:09:48+00:00"
         },
         "publication_package": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:42:48+00:00"
+          "generated_at_utc": "2026-06-18T12:24:04+00:00"
         },
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:43:59+00:00"
+          "generated_at_utc": "2026-06-18T12:24:00+00:00"
         }
       },
       "failures": {}

data/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:10:47+00:00",
+  "generated_at_utc": "2026-06-18T13:02:10+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -215,8 +215,8 @@
     "github_repo": {
       "root": "repo",
       "exists": true,
-      "file_count": 1321,
-      "text_file_count": 1108,
+      "file_count": 1352,
+      "text_file_count": 1129,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -226,8 +226,8 @@
     "hf_space_bundle": {
       "root": "hf_publish/space",
       "exists": true,
-      "file_count": 1103,
-      "text_file_count": 915,
+      "file_count": 1221,
+      "text_file_count": 992,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061
@@ -237,8 +237,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 2582,
-      "text_file_count": 1121,
+      "file_count": 2648,
+      "text_file_count": 1141,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061
@@ -248,8 +248,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 3001,
-      "text_file_count": 1283,
+      "file_count": 3112,
+      "text_file_count": 1309,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061

data/quality_gates.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Release Checks",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "rule": "A release is current when the automated reports pass and the live GitHub/Hugging Face mirrors are verified after publishing.",
   "automated_gates": [
     {

data/qwen3_full_parameter_gates.json

@@ -1,6 +1,6 @@
 {
   "title": "Qwen3-Omni Full-Parameter Feasibility Gates",
-  "generated_at_utc": "2026-06-13T18:14:32+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "status": "pass",
   "decision": "full_parameter_feasible_for_guarded_short_runs_not_promoted",
   "interpretation": "The full-parameter gates prove that Qwen3-Omni full-parameter FSDP can load, prepare, run backward/optimizer steps, and complete guarded pilots up to 256 optimizer steps on an 8-GPU remote worker. They do not prove a production full-parameter fine-tune, and they intentionally save no full checkpoints or public weights.",

data/scope_claims_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:48+00:00",
+  "generated_at_utc": "2026-06-18T12:54:20+00:00",
   "summary": {
     "qwen3_omni_verified_diagnostic_pilot": true,
     "dataset_manifest_num_episodes": 119,

data/single_episode_task_model_radar.json

@@ -1,7 +1,7 @@
 {
   "title": "Single-Episode 20-Task Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "description": "Minimal and Neural MLP baselines on the one public sample episode, both scored on all 20 task contracts.",
   "task_count": 20,
   "method_count": 2,
@@ -13,7 +13,7 @@
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "source_unified_radar": "docs/data/unified_task_model_radar.json",

data/source_alignment_audit.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Source Alignment Note",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:45+00:00",
+  "generated_at_utc": "2026-06-18T12:54:18+00:00",
   "alignment_json": "docs/data/xperience10m_dataset_card_alignment.json",
   "alignment_summary": {
     "full_dataset_repo": "ropedia-ai/xperience-10m",

data/task_method_20_gap_audit.json

@@ -1,10 +1,10 @@
 {
-  "generated_at_utc": "2026-06-18T12:07:14+00:00",
+  "generated_at_utc": "2026-06-18T12:52:47+00:00",
   "immediate_actions": [
     {
       "artifact": "docs/data/task_method_20_gap_audit.json",
       "id": "gap_audit",
-      "purpose": "Keep the 53 scoreless cells visible and reproducible."
+      "purpose": "Keep the 37 scoreless cells visible and reproducible."
     },
     {
       "artifact": "scripts/omni/score_model_output_probes.py",
@@ -45,30 +45,29 @@
       }
     },
     "metadata128_neural_mlp": {
-      "kind": "partial_128_episode_metadata_baseline",
-      "label": "128ep Metadata NN",
+      "kind": "partial_128_episode_aligned_baseline",
+      "label": "128ep Aligned NN",
       "proxy_scored_task_count": 0,
       "result_record_count": 20,
-      "scope": "128 selected episodes, JSONL metadata/text only",
-      "scored_task_count": 7,
-      "scoreless_task_count": 13,
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
+      "scored_task_count": 18,
+      "scoreless_task_count": 2,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 7,
-        "unsupported_without_required_target": 6
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       }
     },
     "metadata128_simple": {
-      "kind": "partial_128_episode_metadata_baseline",
-      "label": "128ep Metadata Simple",
+      "kind": "partial_128_episode_aligned_baseline",
+      "label": "128ep Aligned Simple",
       "proxy_scored_task_count": 0,
       "result_record_count": 20,
-      "scope": "128 selected episodes, JSONL metadata/text only",
-      "scored_task_count": 13,
-      "scoreless_task_count": 7,
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
+      "scored_task_count": 18,
+      "scoreless_task_count": 2,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       }
     },
     "minimal": {
@@ -138,31 +137,22 @@
   "missing_by_method": {
     "cosmos3_nano_future_window": 15,
     "cosmos3_super_reasoner": 13,
-    "metadata128_neural_mlp": 13,
-    "metadata128_simple": 7,
+    "metadata128_neural_mlp": 2,
+    "metadata128_simple": 2,
     "qwen3_omni_v6_lora": 5
   },
   "missing_by_status": {
     "not_evaluated_in_verified_package": 33,
-    "not_supported_by_metadata_only_package": 7,
-    "unsupported_without_required_target": 13
+    "not_supported_by_metadata_only_package": 2,
+    "unsupported_without_required_target": 2
   },
   "missing_by_task": {
-    "01 Action Recognition": [
-      "metadata128_neural_mlp"
-    ],
     "02 Procedure Step Recognition": [
-      "cosmos3_nano_future_window",
-      "metadata128_neural_mlp"
-    ],
-    "04 Next-Action Prediction": [
-      "metadata128_neural_mlp"
+      "cosmos3_nano_future_window"
     ],
     "05 Hand Trajectory Forecasting": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "07 Object Relevance Prediction": [
@@ -173,15 +163,11 @@
       "cosmos3_super_reasoner"
     ],
     "09 Cross-Modal Retrieval": [
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple"
+      "cosmos3_super_reasoner"
     ],
     "10 Cross-Modal Reconstruction": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "11 Temporal Order Verification": [
@@ -190,19 +176,15 @@
     ],
     "12 Multimodal Synchronization Detection": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple"
+      "cosmos3_super_reasoner"
     ],
     "13 Long-Horizon Next-Action Forecasting": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp"
+      "cosmos3_super_reasoner"
     ],
     "14 Long-Horizon Next-Subtask Forecasting": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp"
+      "cosmos3_super_reasoner"
     ],
     "15 Interaction Text Prediction": [
       "cosmos3_nano_future_window",
@@ -212,8 +194,7 @@
       "qwen3_omni_v6_lora"
     ],
     "16 Action-Object Relation Prediction": [
-      "cosmos3_nano_future_window",
-      "metadata128_neural_mlp"
+      "cosmos3_nano_future_window"
     ],
     "17 Future Object-Set Forecasting": [
       "cosmos3_nano_future_window",
@@ -222,8 +203,6 @@
     "18 IMU-to-Hand Pose Reconstruction": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "19 Camera-View Synchronization Retrieval": [
@@ -239,32 +218,6 @@
     ]
   },
   "missing_records": [
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 896 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "timeline_action",
-      "task_label": "Action Recognition",
-      "task_number": 1
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 652 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "timeline_subtask",
-      "task_label": "Procedure Step Recognition",
-      "task_number": 2
-    },
     {
       "method": "Cosmos3-Nano Future Window",
       "metric_key": "macro_f1",
@@ -278,45 +231,6 @@
       "task_label": "Procedure Step Recognition",
       "task_number": 2
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 891 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "next_action",
-      "task_label": "Next-Action Prediction",
-      "task_number": 4
-    },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mpjpe",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "hand_trajectory_forecast",
-      "task_label": "Hand Trajectory Forecasting",
-      "task_number": 5
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mpjpe",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "hand_trajectory_forecast",
-      "task_label": "Hand Trajectory Forecasting",
-      "task_number": 5
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "mpjpe",
@@ -395,32 +309,6 @@
       "task_label": "Language Grounding",
       "task_number": 8
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mrr",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "cross_modal_retrieval",
-      "task_label": "Cross-Modal Retrieval",
-      "task_number": 9
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mrr",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "cross_modal_retrieval",
-      "task_label": "Cross-Modal Retrieval",
-      "task_number": 9
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "mrr",
@@ -434,32 +322,6 @@
       "task_label": "Cross-Modal Retrieval",
       "task_number": 9
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "r2",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "modality_reconstruction",
-      "task_label": "Cross-Modal Reconstruction",
-      "task_number": 10
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "r2",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "modality_reconstruction",
-      "task_label": "Cross-Modal Reconstruction",
-      "task_number": 10
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "r2",
@@ -525,32 +387,6 @@
       "task_label": "Temporal Order Verification",
       "task_number": 11
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "f1",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "misalignment_detection",
-      "task_label": "Multimodal Synchronization Detection",
-      "task_number": 12
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "f1",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "misalignment_detection",
-      "task_label": "Multimodal Synchronization Detection",
-      "task_number": 12
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "f1",
@@ -577,19 +413,6 @@
       "task_label": "Multimodal Synchronization Detection",
       "task_number": 12
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 887 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "long_horizon_next_action",
-      "task_label": "Long-Horizon Next-Action Forecasting",
-      "task_number": 13
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "macro_f1",
@@ -616,19 +439,6 @@
       "task_label": "Long-Horizon Next-Action Forecasting",
       "task_number": 13
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 651 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "next_subtask_forecast",
-      "task_label": "Long-Horizon Next-Subtask Forecasting",
-      "task_number": 14
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "macro_f1",
@@ -656,11 +466,11 @@
       "task_number": 14
     },
     {
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "metric_key": "macro_f1",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
       "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
@@ -669,11 +479,11 @@
       "task_number": 15
     },
     {
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "metric_key": "macro_f1",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
       "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_neural_mlp",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
@@ -720,19 +530,6 @@
       "task_label": "Interaction Text Prediction",
       "task_number": 15
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 3058 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "action_object_relation",
-      "task_label": "Action-Object Relation Prediction",
-      "task_number": 16
-    },
     {
       "method": "Cosmos3-Nano Future Window",
       "metric_key": "macro_f1",
@@ -772,32 +569,6 @@
       "task_label": "Future Object-Set Forecasting",
       "task_number": 17
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mae",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "imu_to_hand_pose",
-      "task_label": "IMU-to-Hand Pose Reconstruction",
-      "task_number": 18
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mae",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "imu_to_hand_pose",
-      "task_label": "IMU-to-Hand Pose Reconstruction",
-      "task_number": 18
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "mae",
@@ -838,11 +609,11 @@
       "task_number": 18
     },
     {
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "metric_key": "mrr",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
       "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
@@ -851,11 +622,11 @@
       "task_number": 19
     },
     {
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "metric_key": "mrr",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
       "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_neural_mlp",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
@@ -975,8 +746,8 @@
     "method_count": 9,
     "method_task_record_count": 180,
     "proxy_scored_method_task_count": 4,
-    "scored_method_task_count": 127,
-    "scoreless_method_task_count": 53,
+    "scored_method_task_count": 143,
+    "scoreless_method_task_count": 37,
     "task_count": 20
   },
   "source_matrix": "docs/data/task_method_20_result_matrix.json",

data/task_method_20_result_matrix.json

@@ -1,11 +1,11 @@
 {
   "title": "Task Method 20-Result Matrix",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "task_count": 20,
   "method_count": 9,
   "method_task_record_count": 180,
-  "scored_method_task_count": 133,
+  "scored_method_task_count": 143,
   "series": [
     {
       "id": "minimal",
@@ -55,50 +55,50 @@
     },
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -264,7 +264,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -274,7 +274,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -282,7 +282,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -292,7 +292,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -426,7 +426,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -436,7 +436,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -444,7 +444,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -454,7 +454,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -588,7 +588,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -598,7 +598,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -606,7 +606,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -616,7 +616,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -750,7 +750,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -760,7 +760,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -768,7 +768,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -778,7 +778,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -912,36 +912,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -1074,7 +1074,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1084,7 +1084,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1092,7 +1092,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1102,7 +1102,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1236,7 +1236,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1246,7 +1246,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1254,7 +1254,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1264,7 +1264,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1398,7 +1398,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1408,7 +1408,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1416,7 +1416,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1426,7 +1426,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1560,36 +1560,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -1722,36 +1722,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -1884,7 +1884,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1894,7 +1894,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1902,7 +1902,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1912,7 +1912,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2046,36 +2046,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -2208,7 +2208,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2218,7 +2218,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2226,7 +2226,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2236,7 +2236,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2370,7 +2370,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2380,7 +2380,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2388,7 +2388,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2398,7 +2398,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2532,7 +2532,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -2542,7 +2542,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -2550,7 +2550,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -2560,8 +2560,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -2694,7 +2694,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2704,7 +2704,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2712,7 +2712,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2722,7 +2722,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2856,7 +2856,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2866,7 +2866,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2874,7 +2874,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2884,7 +2884,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3018,36 +3018,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -3180,7 +3180,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -3190,7 +3190,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -3198,7 +3198,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -3208,8 +3208,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -3342,7 +3342,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3352,7 +3352,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3360,7 +3360,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3370,7 +3370,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

data/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:25+00:00",
+  "generated_at_utc": "2026-06-18T12:54:18+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,

data/unified_task_model_radar.json

@@ -1,18 +1,18 @@
 {
   "title": "Unified 20-Task Model Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "task_count": 20,
   "method_count": 9,
   "method_task_record_count": 180,
-  "scored_method_task_count": 133,
+  "scored_method_task_count": 143,
   "normalization_policy": {
     "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
     "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "series": [
@@ -64,50 +64,50 @@
     },
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -301,7 +301,7 @@
           "raw": 0.008252821966746326,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008252821966746326,
@@ -312,7 +312,7 @@
           "raw": 0.004175793689174209,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004175793689174209,
@@ -401,7 +401,7 @@
           "raw": 0.00019512195121951218,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.00019512195121951218,
@@ -412,7 +412,7 @@
           "raw": 7.207207207207208e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 7.207207207207208e-05,
@@ -523,7 +523,7 @@
           "raw": 0.29652162550029315,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.29652162550029315,
@@ -534,7 +534,7 @@
           "raw": 0.4841733292368365,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4841733292368365,
@@ -634,7 +634,7 @@
           "raw": 0.006514774539765508,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.006514774539765508,
@@ -645,7 +645,7 @@
           "raw": 0.004910507980164745,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004910507980164745,
@@ -709,15 +709,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 8.817333221435547,
           "metric_key": "mpjpe",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.012231610603598841,
+          "raw_text": "8.817",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.429434210062027,
+          "metric_key": "mpjpe",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.25114484128127007,
+          "raw_text": "0.4294",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.2729249894618988,
@@ -741,17 +752,6 @@
           "raw_text": "0.1848",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mpjpe",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "mpjpe",
@@ -856,7 +856,7 @@
           "raw": 0.4381481308057444,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4381481308057444,
@@ -867,7 +867,7 @@
           "raw": 0.5682695682695682,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.5682695682695682,
@@ -956,7 +956,7 @@
           "raw": 0.17764578833693304,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17764578833693304,
@@ -967,7 +967,7 @@
           "raw": 0.18662723837686876,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.18662723837686876,
@@ -1056,7 +1056,7 @@
           "raw": 0.002332374220713973,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.002332374220713973,
@@ -1067,7 +1067,7 @@
           "raw": 0.008236799389123917,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008236799389123917,
@@ -1175,15 +1175,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.002587692579254508,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.002587692579254508,
+          "raw_text": "0.0026",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.0026067993603646755,
+          "metric_key": "mrr",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0026067993603646755,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.003459817497059703,
@@ -1207,17 +1218,6 @@
           "raw_text": "0.0025",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mrr",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "cosmos3_super_reasoner": {
           "raw": null,
           "metric_key": "mrr",
@@ -1264,15 +1264,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": -190.66106203944798,
           "metric_key": "r2",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-190.66",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": -0.43481132003942147,
+          "metric_key": "r2",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-0.4348",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": -1.3450960391924882,
@@ -1296,17 +1307,6 @@
           "raw_text": "-1.397",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "r2",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "r2",
@@ -1389,7 +1389,7 @@
           "raw": 0.4198864140782312,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4198864140782312,
@@ -1400,7 +1400,7 @@
           "raw": 0.8252408266656923,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.8252408266656923,
@@ -1497,15 +1497,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.49980060227663614,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.49980060227663614,
+          "raw_text": "0.4998",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.7773773780941162,
+          "metric_key": "f1",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.7773773780941162,
+          "raw_text": "0.7774",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.4958867673901769,
@@ -1529,17 +1540,6 @@
           "raw_text": "0.8273",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "f1",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "cosmos3_super_reasoner": {
           "raw": null,
           "metric_key": "f1",
@@ -1611,7 +1611,7 @@
           "raw": 0.004579592783699693,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004579592783699693,
@@ -1622,7 +1622,7 @@
           "raw": 0.0029821307969142615,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0029821307969142615,
@@ -1722,7 +1722,7 @@
           "raw": 0.0001206030150753769,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0001206030150753769,
@@ -1733,7 +1733,7 @@
           "raw": 2.086049543676662e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 2.086049543676662e-05,
@@ -1822,7 +1822,7 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
           "normalized_score": null,
@@ -1855,9 +1855,9 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1955,7 +1955,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1966,7 +1966,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -2055,7 +2055,7 @@
           "raw": 0.17656983343047333,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17656983343047333,
@@ -2066,7 +2066,7 @@
           "raw": 0.17418550827844048,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17418550827844048,
@@ -2152,15 +2152,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.2294670194387436,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.18324815505876868,
+          "raw_text": "0.2295",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.2555866539478302,
+          "metric_key": "mae",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.16452114110609004,
+          "raw_text": "0.2556",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.22941437363624573,
@@ -2184,17 +2195,6 @@
           "raw_text": "0.2530",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mae",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "mae",
@@ -2266,7 +2266,7 @@
           "raw": null,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
           "normalized_score": null,
@@ -2299,9 +2299,9 @@
           "raw": null,
           "metric_key": "mrr",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -2388,7 +2388,7 @@
           "raw": 624.8108520507812,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.016864874132806403,
@@ -2399,7 +2399,7 @@
           "raw": 41.4664421081543,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.25411768748242325,
@@ -2456,18 +2456,18 @@
   "model_branch_cards": [
     {
       "id": "metadata128_simple",
-      "title": "128ep Metadata Simple",
+      "title": "128ep Aligned Simple",
       "status": "a100_rerun_pass",
-      "coverage": "20 records / 13 scored JSONL-supported axes",
+      "coverage": "20 records / 18 scored aligned axes",
       "headline": "34,269 rows; train/val/test 25,629/4,608/4,032",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/summary_report.json"
     },
     {
       "id": "metadata128_neural_mlp",
-      "title": "128ep Metadata NN",
+      "title": "128ep Aligned NN",
       "status": "a100_rerun_pass",
-      "coverage": "20 records / 13 scored JSONL-supported axes",
-      "headline": "compact MLP heads over metadata/text features",
+      "coverage": "20 records / 18 scored aligned axes",
+      "headline": "compact MLP heads over metadata/text and staged block features",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/summary_report.json"
     },
     {
@@ -2562,7 +2562,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2572,7 +2572,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2580,7 +2580,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2590,7 +2590,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2724,7 +2724,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2734,7 +2734,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2742,7 +2742,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2752,7 +2752,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2886,7 +2886,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2896,7 +2896,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2904,7 +2904,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2914,7 +2914,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3048,7 +3048,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3058,7 +3058,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3066,7 +3066,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3076,7 +3076,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3210,36 +3210,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -3372,7 +3372,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3382,7 +3382,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3390,7 +3390,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3400,7 +3400,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3534,7 +3534,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3544,7 +3544,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3552,7 +3552,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3562,7 +3562,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3696,7 +3696,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3706,7 +3706,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3714,7 +3714,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3724,7 +3724,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3858,36 +3858,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -4020,36 +4020,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -4182,7 +4182,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4192,7 +4192,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4200,7 +4200,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4210,7 +4210,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4344,36 +4344,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -4506,7 +4506,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4516,7 +4516,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4524,7 +4524,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4534,7 +4534,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4668,7 +4668,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4678,7 +4678,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4686,7 +4686,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4696,7 +4696,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4830,7 +4830,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -4840,7 +4840,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -4848,7 +4848,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -4858,8 +4858,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -4992,7 +4992,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5002,7 +5002,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5010,7 +5010,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5020,7 +5020,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5154,7 +5154,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5164,7 +5164,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5172,7 +5172,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5182,7 +5182,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5316,36 +5316,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -5478,7 +5478,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -5488,7 +5488,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -5496,7 +5496,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -5506,8 +5506,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -5640,7 +5640,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5650,7 +5650,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5658,7 +5658,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5668,7 +5668,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

data/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:46+00:00",
+  "generated_at_utc": "2026-06-18T12:54:19+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 116110,
+      "bytes": 116111,
       "top_level_type": "dict"
     },
     {
@@ -316,7 +316,7 @@
     },
     {
       "path": "data/episode128_task_model_radar.json",
-      "bytes": 186443,
+      "bytes": 185447,
       "top_level_type": "dict"
     },
     {
@@ -351,7 +351,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 994053,
+      "bytes": 1059014,
       "top_level_type": "dict"
     },
     {
@@ -471,7 +471,7 @@
     },
     {
       "path": "data/single_episode_task_model_radar.json",
-      "bytes": 50973,
+      "bytes": 51064,
       "top_level_type": "dict"
     },
     {
@@ -486,12 +486,12 @@
     },
     {
       "path": "data/task_method_20_gap_audit.json",
-      "bytes": 46902,
+      "bytes": 35883,
       "top_level_type": "dict"
     },
     {
       "path": "data/task_method_20_result_matrix.json",
-      "bytes": 129242,
+      "bytes": 128794,
       "top_level_type": "dict"
     },
     {
@@ -526,7 +526,7 @@
     },
     {
       "path": "data/unified_task_model_radar.json",
-      "bytes": 230297,
+      "bytes": 229299,
       "top_level_type": "dict"
     },
     {
@@ -571,7 +571,7 @@
     {
       "path": "assets/charts/episode128_task_model_radar.svg",
       "exists": true,
-      "bytes": 45937,
+      "bytes": 47540,
       "format": "SVG",
       "has_viewbox": true
     },
@@ -641,7 +641,7 @@
     {
       "path": "assets/charts/unified_task_model_radar.svg",
       "exists": true,
-      "bytes": 51953,
+      "bytes": 53553,
       "format": "SVG",
       "has_viewbox": true
     },
@@ -752,7 +752,7 @@
     {
       "path": "assets/task_suite_infographic.png",
       "exists": true,
-      "bytes": 2627286,
+      "bytes": 1591194,
       "width": 1800,
       "height": 6600,
       "format": "PNG"

docs/data/episode128_task_model_radar.json

@@ -1,19 +1,19 @@
 {
   "title": "128-Episode 20-Task Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "description": "Selected 128-episode metadata/raw baselines plus verified Qwen3/Cosmos branches. Every method has 20 records; numeric scores appear only where the public artifact produced that task target.",
   "task_count": 20,
   "method_count": 7,
   "method_task_record_count": 140,
-  "scored_method_task_count": 93,
+  "scored_method_task_count": 103,
   "normalization_policy": {
     "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
     "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "source_unified_radar": "docs/data/unified_task_model_radar.json",
@@ -21,50 +21,50 @@
   "series": [
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -205,7 +205,7 @@
           "raw": 0.008252821966746326,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008252821966746326,
@@ -216,7 +216,7 @@
           "raw": 0.004175793689174209,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004175793689174209,
@@ -296,7 +296,7 @@
           "raw": 0.00019512195121951218,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.00019512195121951218,
@@ -307,7 +307,7 @@
           "raw": 7.207207207207208e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 7.207207207207208e-05,
@@ -387,7 +387,7 @@
           "raw": 0.29652162550029315,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.29652162550029315,
@@ -398,7 +398,7 @@
           "raw": 0.4841733292368365,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4841733292368365,
@@ -478,7 +478,7 @@
           "raw": 0.006514774539765508,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.006514774539765508,
@@ -489,7 +489,7 @@
           "raw": 0.004910507980164745,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004910507980164745,
@@ -566,26 +566,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 8.817333221435547,
           "metric_key": "mpjpe",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.012231610603598841,
+          "raw_text": "8.817",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.429434210062027,
           "metric_key": "mpjpe",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.25114484128127007,
+          "raw_text": "0.4294",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.2729249894618988,
@@ -660,7 +660,7 @@
           "raw": 0.4381481308057444,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4381481308057444,
@@ -671,7 +671,7 @@
           "raw": 0.5682695682695682,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.5682695682695682,
@@ -751,7 +751,7 @@
           "raw": 0.17764578833693304,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17764578833693304,
@@ -762,7 +762,7 @@
           "raw": 0.18662723837686876,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.18662723837686876,
@@ -842,7 +842,7 @@
           "raw": 0.002332374220713973,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.002332374220713973,
@@ -853,7 +853,7 @@
           "raw": 0.008236799389123917,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008236799389123917,
@@ -930,26 +930,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.002587692579254508,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.002587692579254508,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.0026067993603646755,
           "metric_key": "mrr",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0026067993603646755,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.003459817497059703,
@@ -1021,26 +1021,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": -190.66106203944798,
           "metric_key": "r2",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-190.66",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": -0.43481132003942147,
           "metric_key": "r2",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-0.4348",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": -1.3450960391924882,
@@ -1115,7 +1115,7 @@
           "raw": 0.4198864140782312,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4198864140782312,
@@ -1126,7 +1126,7 @@
           "raw": 0.8252408266656923,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.8252408266656923,
@@ -1203,26 +1203,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.49980060227663614,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.49980060227663614,
+          "raw_text": "0.4998",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.7773773780941162,
           "metric_key": "f1",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.7773773780941162,
+          "raw_text": "0.7774",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.4958867673901769,
@@ -1297,7 +1297,7 @@
           "raw": 0.004579592783699693,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004579592783699693,
@@ -1308,7 +1308,7 @@
           "raw": 0.0029821307969142615,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0029821307969142615,
@@ -1388,7 +1388,7 @@
           "raw": 0.0001206030150753769,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0001206030150753769,
@@ -1399,7 +1399,7 @@
           "raw": 2.086049543676662e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 2.086049543676662e-05,
@@ -1479,7 +1479,7 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
           "normalized_score": null,
@@ -1490,9 +1490,9 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1570,7 +1570,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1581,7 +1581,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1661,7 +1661,7 @@
           "raw": 0.17656983343047333,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17656983343047333,
@@ -1672,7 +1672,7 @@
           "raw": 0.17418550827844048,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17418550827844048,
@@ -1749,26 +1749,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.2294670194387436,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.18324815505876868,
+          "raw_text": "0.2295",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.2555866539478302,
           "metric_key": "mae",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.16452114110609004,
+          "raw_text": "0.2556",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.22941437363624573,
@@ -1843,7 +1843,7 @@
           "raw": null,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
           "normalized_score": null,
@@ -1854,9 +1854,9 @@
           "raw": null,
           "metric_key": "mrr",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1934,7 +1934,7 @@
           "raw": 624.8108520507812,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.016864874132806403,
@@ -1945,7 +1945,7 @@
           "raw": 41.4664421081543,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.25411768748242325,
@@ -2016,7 +2016,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2026,7 +2026,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2034,7 +2034,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2044,7 +2044,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2142,7 +2142,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2152,7 +2152,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2160,7 +2160,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2170,7 +2170,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2268,7 +2268,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2278,7 +2278,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2286,7 +2286,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2296,7 +2296,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2394,7 +2394,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2404,7 +2404,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2412,7 +2412,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2422,7 +2422,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2520,36 +2520,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -2646,7 +2646,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2656,7 +2656,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2664,7 +2664,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2674,7 +2674,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2772,7 +2772,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2782,7 +2782,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2790,7 +2790,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2800,7 +2800,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2898,7 +2898,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2908,7 +2908,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2916,7 +2916,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2926,7 +2926,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3024,36 +3024,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -3150,36 +3150,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -3276,7 +3276,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3286,7 +3286,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3294,7 +3294,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3304,7 +3304,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3402,36 +3402,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -3528,7 +3528,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3538,7 +3538,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3546,7 +3546,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3556,7 +3556,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3654,7 +3654,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3664,7 +3664,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3672,7 +3672,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3682,7 +3682,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3780,7 +3780,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -3790,7 +3790,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -3798,7 +3798,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -3808,8 +3808,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -3906,7 +3906,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3916,7 +3916,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3924,7 +3924,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3934,7 +3934,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4032,7 +4032,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4042,7 +4042,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4050,7 +4050,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4060,7 +4060,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4158,36 +4158,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -4284,7 +4284,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -4294,7 +4294,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -4302,7 +4302,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -4312,8 +4312,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -4410,7 +4410,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4420,7 +4420,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4428,7 +4428,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4438,7 +4438,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

docs/data/omni_model_comparison.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Current Result Versions and Model Groups",
-  "generated_at_utc": "2026-06-13T18:14:42+00:00",
+  "generated_at_utc": "2026-06-18T12:52:47+00:00",
   "status": "pass",
   "version_count": 3,
   "model_group_count": 5,

docs/data/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -18,7 +18,7 @@
         "website_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:41:43+00:00"
+          "generated_at_utc": "2026-06-18T12:09:46+00:00"
         },
         "rendered_site_check": {
           "exists": true,
@@ -28,27 +28,27 @@
         "task_surface_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:25+00:00"
         },
         "source_alignment": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:45+00:00"
         },
         "scale_up_status": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:06+00:00"
+          "generated_at_utc": "2026-06-18T12:09:48+00:00"
         },
         "publication_package": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:42:48+00:00"
+          "generated_at_utc": "2026-06-18T12:24:04+00:00"
         },
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:43:59+00:00"
+          "generated_at_utc": "2026-06-18T12:24:00+00:00"
         }
       },
       "failures": {}

docs/data/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:25+00:00",
+  "generated_at_utc": "2026-06-18T12:54:18+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,

metrics/artifact_index.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Task Suite Artifact Index",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:52:48+00:00",
   "status": "pass",
   "artifact_count": 213,
   "missing": [],
@@ -290,8 +290,8 @@
       "surface": "repo_hf",
       "shows": "Runs simple metadata and neural MLP baselines on the same selected 96/16/16 episode split used by the Qwen3-Omni diagnostic pilot.",
       "exists": true,
-      "bytes": 73236,
-      "sha256": "76acae0de25d51413e7e6f11021163e7d9909cfe95d65bf6b02e74043d429e2d"
+      "bytes": 74368,
+      "sha256": "6f54bfb963d5102ebd61eb8f8b6d8f6919db673378c9d5940d89ec5ea6f3d4b2"
     },
     {
       "id": "task_suite_enhancement_128",
@@ -599,7 +599,7 @@
       "shows": "Machine-readable source-alignment pass/fail check for repo, website, and HF surfaces.",
       "exists": true,
       "bytes": 4432,
-      "sha256": "ae089cc0df132b63365e03b2157a488b5d1569567c0374d7621bcd347da62c9e"
+      "sha256": "8ddadfe15ba8779e82879f965ff50bceb9c573bc942c3ecf176fbf20e5faeaea"
     },
     {
       "id": "source_alignment_validator",
@@ -719,8 +719,8 @@
       "surface": "website_hf",
       "shows": "Stores normalized 20-axis radar values, raw task metrics, Qwen3/Cosmos overlay mappings, branch-card caveats, and explicit scoreless status records.",
       "exists": true,
-      "bytes": 230297,
-      "sha256": "437874b1633e73165e3300f55580394663a44759c848288e696859b98f8aad32"
+      "bytes": 229299,
+      "sha256": "30f338139df391c36941da0b759cc237366ee43d006bfff2d2e43481cc2d2a63"
     },
     {
       "id": "single_episode_task_model_radar_json",
@@ -730,8 +730,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable split radar for the one-episode Minimal and Neural MLP baselines, both scored on all 20 task contracts.",
       "exists": true,
-      "bytes": 50973,
-      "sha256": "38cb43512f2ac40feeb62333bdea89b3a55e5b48468beb8982cf22536f794ecf"
+      "bytes": 51064,
+      "sha256": "52001c8ac081b14827a8a55cae21da8fd32516f81365d7dda1047ef68096eef8"
     },
     {
       "id": "episode128_task_model_radar_json",
@@ -741,8 +741,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable split radar for selected 128-episode metadata/raw baselines and verified Qwen3/Cosmos branches, preserving explicit scoreless cells.",
       "exists": true,
-      "bytes": 186443,
-      "sha256": "55e758e8703f406889022976d0ba055181212305c9a7246e899463e0c3c3b554"
+      "bytes": 185447,
+      "sha256": "e9994f42a1e086411748e1233761c84a8dcd564898c216454a8872c2f4d4f213"
     },
     {
       "id": "task_method_20_result_matrix_json",
@@ -752,8 +752,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable 9-method by 20-task matrix where every method has 20 records and scoreless cells carry unsupported/not-evaluated reasons.",
       "exists": true,
-      "bytes": 129242,
-      "sha256": "64fb700d51f536edf11291799b6173cf9ae8dd7a41178aac348b8207ed4b1e42"
+      "bytes": 128794,
+      "sha256": "1bce6001518b314fc8a5e86eab56521aa9718d09d787765d10caee4d791e9809"
     },
     {
       "id": "task_method_20_result_matrix",
@@ -763,8 +763,8 @@
       "surface": "repo_hf",
       "shows": "Reader-facing table that separates 20 records per method from numeric scored axes, documented raw128 proxy scores, unsupported metadata targets, and model targets not evaluated in verified packages.",
       "exists": true,
-      "bytes": 4026,
-      "sha256": "55e949fc30419a52f7f5ec4dd9544a11b253b076f8e3637ec3e92b3d61a89aab"
+      "bytes": 3954,
+      "sha256": "01b21d83954f700e4b061e96b1f58c6af474d79a2caaff1bfcff4854b66722ca"
     },
     {
       "id": "task_method_20_gap_audit_json",
@@ -774,8 +774,8 @@
       "surface": "website_hf",
       "shows": "Machine-readable 180-record gap ledger with numeric scores, scoreless cells, explicit status reasons, and next evidence needed before new scores can be published.",
       "exists": true,
-      "bytes": 46902,
-      "sha256": "2b64dbd013625852679f9b91d25c48d1ed197fec727883b4fe37088b2d594784"
+      "bytes": 35883,
+      "sha256": "9336756d67d2488a28c4bb9c282f65230031eeb8dddd087a11fd441d8e61539b"
     },
     {
       "id": "task_method_20_gap_audit",
@@ -785,8 +785,8 @@
       "surface": "repo_hf",
       "shows": "Reader-facing ledger that lists every scoreless method-task cell and the concrete target or model-output evidence required before it can become numeric.",
       "exists": true,
-      "bytes": 13387,
-      "sha256": "d33461eb704f8e92545b6b54d9fc509e617fbacc9ca9894ac851ca9c3dec0fec"
+      "bytes": 10286,
+      "sha256": "45969b72e9a3ff8c40d958ea819e725fd4df5d90424ccdffd1c64fd1a5152063"
     },
     {
       "id": "unified_task_model_radar_chart",
@@ -796,8 +796,8 @@
       "surface": "website_hf",
       "shows": "Compares minimal and neural MLP baselines across all 20 tasks, with Qwen3/Cosmos task-aligned model overlays.",
       "exists": true,
-      "bytes": 51953,
-      "sha256": "19c001f10319946ef0e4921064f8a012836f29e7c8b272f900c257169faf46a1"
+      "bytes": 53553,
+      "sha256": "ec9a8bf0f5814106ddb8e62d0941c7cc07d1b8a29323a61a400319ffe6bd3485"
     },
     {
       "id": "single_episode_task_model_radar_chart",
@@ -818,8 +818,8 @@
       "surface": "website_hf",
       "shows": "Separates the selected 128-episode methods: raw-feature simple/NN as complete 20/20 scored polygons and metadata/Qwen/Cosmos as task-aligned overlays.",
       "exists": true,
-      "bytes": 45937,
-      "sha256": "b504b1b9c5cad0caa8c822d5bb2971c1b708251cf7b9ef587a92db2c12751e97"
+      "bytes": 47540,
+      "sha256": "0c2283a04fe401851b8b313de3ba383d24185262f4c6500d12fa0a3b8c0c4443"
     },
     {
       "id": "unified_task_model_radar_builder",
@@ -829,8 +829,8 @@
       "surface": "repo_hf",
       "shows": "Regenerates the direction-aware radar chart and machine-readable metric overlay JSON.",
       "exists": true,
-      "bytes": 52388,
-      "sha256": "f4803360cfd02383a1942a93a5845308db936b479a5b906719e46e192f3ef142"
+      "bytes": 52743,
+      "sha256": "e081f88e9f31934b24820c5cbffb957bb235a3275f553e573ab44e5c3d03c99a"
     },
     {
       "id": "task_method_20_gap_audit_builder",
@@ -906,8 +906,8 @@
       "surface": "repo_hf",
       "shows": "Rerun of JSONL metadata/text simple and neural baselines over the selected 128-episode multiscale dataset; supports radar overlays on JSONL-supported task axes.",
       "exists": true,
-      "bytes": 109248,
-      "sha256": "5e7f3085be5012eb3dda46f9c7b5b7c0ae22d6a0fbce71d6e99dd317fecc12af"
+      "bytes": 124232,
+      "sha256": "dba221a6ed8a6a84602dc21a1055cbb4444c03775f74b55e5d72861941820ac8"
     },
     {
       "id": "a100_128_raw20_task_baselines",
@@ -1310,7 +1310,7 @@
       "volatile": true,
       "shows": "Confirms prepared GitHub/HF Space/artifact/model mirrors share the same critical data, figure, website HTML, and validator files.",
       "exists": true,
-      "bytes": 994053,
+      "bytes": 1059014,
       "hash_policy": "existence_and_size_only"
     },
     {
@@ -1620,7 +1620,7 @@
       "shows": "Reader-facing comparison of the single-episode task suite, 128-episode aligned baselines, Qwen3-Omni packages, and Cosmos3 future-window branch.",
       "exists": true,
       "bytes": 15999,
-      "sha256": "30053bdea6c417ab02f98d99d8e80cd7e304bc3a9dfacbf599139d3221c02c8f"
+      "sha256": "dd65ae9077acbce91870b182d701db367a9c79eb287aeee2a1e165ec4915e5f3"
     },
     {
       "id": "omni_model_comparison_json",
@@ -1631,7 +1631,7 @@
       "shows": "Machine-readable comparison of the current result versions, per-task aligned baselines, verified Qwen3 packages, and Cosmos3 package.",
       "exists": true,
       "bytes": 81866,
-      "sha256": "1c9d4ba370661b0e0cb7104e9a51abdc3fe91a440ae86e748b10b719d1d613cc"
+      "sha256": "dd7a599117defcc1fd783c3134b6b3fc92f2ec2190ea517624cb215b931bd87a"
     },
     {
       "id": "cosmos3_nano_verified_summary",

metrics/episode128_task_model_radar.json

@@ -1,19 +1,19 @@
 {
   "title": "128-Episode 20-Task Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "description": "Selected 128-episode metadata/raw baselines plus verified Qwen3/Cosmos branches. Every method has 20 records; numeric scores appear only where the public artifact produced that task target.",
   "task_count": 20,
   "method_count": 7,
   "method_task_record_count": 140,
-  "scored_method_task_count": 93,
+  "scored_method_task_count": 103,
   "normalization_policy": {
     "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
     "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "source_unified_radar": "docs/data/unified_task_model_radar.json",
@@ -21,50 +21,50 @@
   "series": [
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -205,7 +205,7 @@
           "raw": 0.008252821966746326,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008252821966746326,
@@ -216,7 +216,7 @@
           "raw": 0.004175793689174209,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004175793689174209,
@@ -296,7 +296,7 @@
           "raw": 0.00019512195121951218,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.00019512195121951218,
@@ -307,7 +307,7 @@
           "raw": 7.207207207207208e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 7.207207207207208e-05,
@@ -387,7 +387,7 @@
           "raw": 0.29652162550029315,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.29652162550029315,
@@ -398,7 +398,7 @@
           "raw": 0.4841733292368365,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4841733292368365,
@@ -478,7 +478,7 @@
           "raw": 0.006514774539765508,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.006514774539765508,
@@ -489,7 +489,7 @@
           "raw": 0.004910507980164745,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004910507980164745,
@@ -566,26 +566,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 8.817333221435547,
           "metric_key": "mpjpe",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.012231610603598841,
+          "raw_text": "8.817",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.429434210062027,
           "metric_key": "mpjpe",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.25114484128127007,
+          "raw_text": "0.4294",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.2729249894618988,
@@ -660,7 +660,7 @@
           "raw": 0.4381481308057444,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4381481308057444,
@@ -671,7 +671,7 @@
           "raw": 0.5682695682695682,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.5682695682695682,
@@ -751,7 +751,7 @@
           "raw": 0.17764578833693304,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17764578833693304,
@@ -762,7 +762,7 @@
           "raw": 0.18662723837686876,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.18662723837686876,
@@ -842,7 +842,7 @@
           "raw": 0.002332374220713973,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.002332374220713973,
@@ -853,7 +853,7 @@
           "raw": 0.008236799389123917,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008236799389123917,
@@ -930,26 +930,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.002587692579254508,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.002587692579254508,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.0026067993603646755,
           "metric_key": "mrr",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0026067993603646755,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.003459817497059703,
@@ -1021,26 +1021,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": -190.66106203944798,
           "metric_key": "r2",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-190.66",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": -0.43481132003942147,
           "metric_key": "r2",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-0.4348",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": -1.3450960391924882,
@@ -1115,7 +1115,7 @@
           "raw": 0.4198864140782312,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4198864140782312,
@@ -1126,7 +1126,7 @@
           "raw": 0.8252408266656923,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.8252408266656923,
@@ -1203,26 +1203,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.49980060227663614,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.49980060227663614,
+          "raw_text": "0.4998",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.7773773780941162,
           "metric_key": "f1",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.7773773780941162,
+          "raw_text": "0.7774",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.4958867673901769,
@@ -1297,7 +1297,7 @@
           "raw": 0.004579592783699693,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004579592783699693,
@@ -1308,7 +1308,7 @@
           "raw": 0.0029821307969142615,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0029821307969142615,
@@ -1388,7 +1388,7 @@
           "raw": 0.0001206030150753769,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0001206030150753769,
@@ -1399,7 +1399,7 @@
           "raw": 2.086049543676662e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 2.086049543676662e-05,
@@ -1479,7 +1479,7 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
           "normalized_score": null,
@@ -1490,9 +1490,9 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1570,7 +1570,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1581,7 +1581,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1661,7 +1661,7 @@
           "raw": 0.17656983343047333,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17656983343047333,
@@ -1672,7 +1672,7 @@
           "raw": 0.17418550827844048,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17418550827844048,
@@ -1749,26 +1749,26 @@
       "raw128_proxy_axis": false,
       "values": {
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.2294670194387436,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.18324815505876868,
+          "raw_text": "0.2295",
+          "status_label": "scored"
         },
         "metadata128_neural_mlp": {
-          "raw": null,
+          "raw": 0.2555866539478302,
           "metric_key": "mae",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.16452114110609004,
+          "raw_text": "0.2556",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.22941437363624573,
@@ -1843,7 +1843,7 @@
           "raw": null,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
           "normalized_score": null,
@@ -1854,9 +1854,9 @@
           "raw": null,
           "metric_key": "mrr",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1934,7 +1934,7 @@
           "raw": 624.8108520507812,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.016864874132806403,
@@ -1945,7 +1945,7 @@
           "raw": 41.4664421081543,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.25411768748242325,
@@ -2016,7 +2016,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2026,7 +2026,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2034,7 +2034,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2044,7 +2044,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2142,7 +2142,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2152,7 +2152,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2160,7 +2160,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2170,7 +2170,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2268,7 +2268,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2278,7 +2278,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2286,7 +2286,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2296,7 +2296,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2394,7 +2394,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2404,7 +2404,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2412,7 +2412,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2422,7 +2422,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2520,36 +2520,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -2646,7 +2646,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2656,7 +2656,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2664,7 +2664,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2674,7 +2674,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2772,7 +2772,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2782,7 +2782,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2790,7 +2790,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2800,7 +2800,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2898,7 +2898,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2908,7 +2908,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2916,7 +2916,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2926,7 +2926,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3024,36 +3024,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -3150,36 +3150,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -3276,7 +3276,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3286,7 +3286,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3294,7 +3294,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3304,7 +3304,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3402,36 +3402,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -3528,7 +3528,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3538,7 +3538,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3546,7 +3546,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3556,7 +3556,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3654,7 +3654,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3664,7 +3664,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3672,7 +3672,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3682,7 +3682,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3780,7 +3780,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -3790,7 +3790,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -3798,7 +3798,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -3808,8 +3808,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -3906,7 +3906,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3916,7 +3916,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3924,7 +3924,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3934,7 +3934,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4032,7 +4032,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4042,7 +4042,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4050,7 +4050,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4060,7 +4060,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4158,36 +4158,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -4284,7 +4284,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -4294,7 +4294,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -4302,7 +4302,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -4312,8 +4312,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -4410,7 +4410,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4420,7 +4420,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4428,7 +4428,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4438,7 +4438,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

metrics/omni_model_comparison.json

@@ -1,6 +1,6 @@
 {
   "title": "Ropedia Xperience-10M Current Result Versions and Model Groups",
-  "generated_at_utc": "2026-06-13T18:14:42+00:00",
+  "generated_at_utc": "2026-06-18T12:52:47+00:00",
   "status": "pass",
   "version_count": 3,
   "model_group_count": 5,

metrics/public_surface_qa.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Public Project Surface",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
   "checks": [
     {
@@ -18,7 +18,7 @@
         "website_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:41:43+00:00"
+          "generated_at_utc": "2026-06-18T12:09:46+00:00"
         },
         "rendered_site_check": {
           "exists": true,
@@ -28,27 +28,27 @@
         "task_surface_integrity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:25+00:00"
         },
         "source_alignment": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:04+00:00"
+          "generated_at_utc": "2026-06-18T12:09:45+00:00"
         },
         "scale_up_status": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:18:06+00:00"
+          "generated_at_utc": "2026-06-18T12:09:48+00:00"
         },
         "publication_package": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:42:48+00:00"
+          "generated_at_utc": "2026-06-18T12:24:04+00:00"
         },
         "mirror_parity": {
           "exists": true,
           "status": "pass",
-          "generated_at_utc": "2026-06-18T11:43:59+00:00"
+          "generated_at_utc": "2026-06-18T12:24:00+00:00"
         }
       },
       "failures": {}

metrics/publication_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:10:47+00:00",
+  "generated_at_utc": "2026-06-18T13:02:10+00:00",
   "checks": [
     {
       "name": "required_publication_assets_present",
@@ -215,8 +215,8 @@
     "github_repo": {
       "root": "repo",
       "exists": true,
-      "file_count": 1321,
-      "text_file_count": 1108,
+      "file_count": 1352,
+      "text_file_count": 1129,
       "largest_file": {
         "path": "results/episode_task_suite/modality_reconstruction/predictions.npz",
         "bytes": 55702978
@@ -226,8 +226,8 @@
     "hf_space_bundle": {
       "root": "hf_publish/space",
       "exists": true,
-      "file_count": 1103,
-      "text_file_count": 915,
+      "file_count": 1221,
+      "text_file_count": 992,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061
@@ -237,8 +237,8 @@
     "hf_artifact_bundle": {
       "root": "hf_publish/artifacts",
       "exists": true,
-      "file_count": 2582,
-      "text_file_count": 1121,
+      "file_count": 2648,
+      "text_file_count": 1141,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061
@@ -248,8 +248,8 @@
     "hf_model_bundle": {
       "root": "hf_publish/model",
       "exists": true,
-      "file_count": 3001,
-      "text_file_count": 1283,
+      "file_count": 3112,
+      "text_file_count": 1309,
       "largest_file": {
         "path": "results/omni_finetune/xperience10m_128ep_dense_multiscale_hierarchical_v1_20260608/dense_multiscale_windows.jsonl",
         "bytes": 135591061

metrics/quality_gates.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Release Checks",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:24+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "rule": "A release is current when the automated reports pass and the live GitHub/Hugging Face mirrors are verified after publishing.",
   "automated_gates": [
     {

metrics/qwen3_full_parameter_gates.json

@@ -1,6 +1,6 @@
 {
   "title": "Qwen3-Omni Full-Parameter Feasibility Gates",
-  "generated_at_utc": "2026-06-13T18:14:32+00:00",
+  "generated_at_utc": "2026-06-18T12:53:13+00:00",
   "status": "pass",
   "decision": "full_parameter_feasible_for_guarded_short_runs_not_promoted",
   "interpretation": "The full-parameter gates prove that Qwen3-Omni full-parameter FSDP can load, prepare, run backward/optimizer steps, and complete guarded pilots up to 256 optimizer steps on an 8-GPU remote worker. They do not prove a production full-parameter fine-tune, and they intentionally save no full checkpoints or public weights.",

metrics/scope_claims_audit.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:48+00:00",
+  "generated_at_utc": "2026-06-18T12:54:20+00:00",
   "summary": {
     "qwen3_omni_verified_diagnostic_pilot": true,
     "dataset_manifest_num_episodes": 119,

metrics/single_episode_task_model_radar.json

@@ -1,7 +1,7 @@
 {
   "title": "Single-Episode 20-Task Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "description": "Minimal and Neural MLP baselines on the one public sample episode, both scored on all 20 task contracts.",
   "task_count": 20,
   "method_count": 2,
@@ -13,7 +13,7 @@
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "source_unified_radar": "docs/data/unified_task_model_radar.json",

metrics/source_alignment_audit.json

@@ -1,7 +1,7 @@
 {
   "title": "Ropedia Xperience-10M Source Alignment Note",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:45+00:00",
+  "generated_at_utc": "2026-06-18T12:54:18+00:00",
   "alignment_json": "docs/data/xperience10m_dataset_card_alignment.json",
   "alignment_summary": {
     "full_dataset_repo": "ropedia-ai/xperience-10m",

metrics/task_method_20_gap_audit.json

@@ -1,10 +1,10 @@
 {
-  "generated_at_utc": "2026-06-18T12:07:14+00:00",
+  "generated_at_utc": "2026-06-18T12:52:47+00:00",
   "immediate_actions": [
     {
       "artifact": "docs/data/task_method_20_gap_audit.json",
       "id": "gap_audit",
-      "purpose": "Keep the 53 scoreless cells visible and reproducible."
+      "purpose": "Keep the 37 scoreless cells visible and reproducible."
     },
     {
       "artifact": "scripts/omni/score_model_output_probes.py",
@@ -45,30 +45,29 @@
       }
     },
     "metadata128_neural_mlp": {
-      "kind": "partial_128_episode_metadata_baseline",
-      "label": "128ep Metadata NN",
+      "kind": "partial_128_episode_aligned_baseline",
+      "label": "128ep Aligned NN",
       "proxy_scored_task_count": 0,
       "result_record_count": 20,
-      "scope": "128 selected episodes, JSONL metadata/text only",
-      "scored_task_count": 7,
-      "scoreless_task_count": 13,
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
+      "scored_task_count": 18,
+      "scoreless_task_count": 2,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 7,
-        "unsupported_without_required_target": 6
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       }
     },
     "metadata128_simple": {
-      "kind": "partial_128_episode_metadata_baseline",
-      "label": "128ep Metadata Simple",
+      "kind": "partial_128_episode_aligned_baseline",
+      "label": "128ep Aligned Simple",
       "proxy_scored_task_count": 0,
       "result_record_count": 20,
-      "scope": "128 selected episodes, JSONL metadata/text only",
-      "scored_task_count": 13,
-      "scoreless_task_count": 7,
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
+      "scored_task_count": 18,
+      "scoreless_task_count": 2,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       }
     },
     "minimal": {
@@ -138,31 +137,22 @@
   "missing_by_method": {
     "cosmos3_nano_future_window": 15,
     "cosmos3_super_reasoner": 13,
-    "metadata128_neural_mlp": 13,
-    "metadata128_simple": 7,
+    "metadata128_neural_mlp": 2,
+    "metadata128_simple": 2,
     "qwen3_omni_v6_lora": 5
   },
   "missing_by_status": {
     "not_evaluated_in_verified_package": 33,
-    "not_supported_by_metadata_only_package": 7,
-    "unsupported_without_required_target": 13
+    "not_supported_by_metadata_only_package": 2,
+    "unsupported_without_required_target": 2
   },
   "missing_by_task": {
-    "01 Action Recognition": [
-      "metadata128_neural_mlp"
-    ],
     "02 Procedure Step Recognition": [
-      "cosmos3_nano_future_window",
-      "metadata128_neural_mlp"
-    ],
-    "04 Next-Action Prediction": [
-      "metadata128_neural_mlp"
+      "cosmos3_nano_future_window"
     ],
     "05 Hand Trajectory Forecasting": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "07 Object Relevance Prediction": [
@@ -173,15 +163,11 @@
       "cosmos3_super_reasoner"
     ],
     "09 Cross-Modal Retrieval": [
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple"
+      "cosmos3_super_reasoner"
     ],
     "10 Cross-Modal Reconstruction": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "11 Temporal Order Verification": [
@@ -190,19 +176,15 @@
     ],
     "12 Multimodal Synchronization Detection": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple"
+      "cosmos3_super_reasoner"
     ],
     "13 Long-Horizon Next-Action Forecasting": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp"
+      "cosmos3_super_reasoner"
     ],
     "14 Long-Horizon Next-Subtask Forecasting": [
       "cosmos3_nano_future_window",
-      "cosmos3_super_reasoner",
-      "metadata128_neural_mlp"
+      "cosmos3_super_reasoner"
     ],
     "15 Interaction Text Prediction": [
       "cosmos3_nano_future_window",
@@ -212,8 +194,7 @@
       "qwen3_omni_v6_lora"
     ],
     "16 Action-Object Relation Prediction": [
-      "cosmos3_nano_future_window",
-      "metadata128_neural_mlp"
+      "cosmos3_nano_future_window"
     ],
     "17 Future Object-Set Forecasting": [
       "cosmos3_nano_future_window",
@@ -222,8 +203,6 @@
     "18 IMU-to-Hand Pose Reconstruction": [
       "cosmos3_nano_future_window",
       "cosmos3_super_reasoner",
-      "metadata128_neural_mlp",
-      "metadata128_simple",
       "qwen3_omni_v6_lora"
     ],
     "19 Camera-View Synchronization Retrieval": [
@@ -239,32 +218,6 @@
     ]
   },
   "missing_records": [
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 896 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "timeline_action",
-      "task_label": "Action Recognition",
-      "task_number": 1
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 652 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "timeline_subtask",
-      "task_label": "Procedure Step Recognition",
-      "task_number": 2
-    },
     {
       "method": "Cosmos3-Nano Future Window",
       "metric_key": "macro_f1",
@@ -278,45 +231,6 @@
       "task_label": "Procedure Step Recognition",
       "task_number": 2
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 891 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "next_action",
-      "task_label": "Next-Action Prediction",
-      "task_number": 4
-    },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mpjpe",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "hand_trajectory_forecast",
-      "task_label": "Hand Trajectory Forecasting",
-      "task_number": 5
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mpjpe",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "hand_trajectory_forecast",
-      "task_label": "Hand Trajectory Forecasting",
-      "task_number": 5
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "mpjpe",
@@ -395,32 +309,6 @@
       "task_label": "Language Grounding",
       "task_number": 8
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mrr",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "cross_modal_retrieval",
-      "task_label": "Cross-Modal Retrieval",
-      "task_number": 9
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mrr",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "cross_modal_retrieval",
-      "task_label": "Cross-Modal Retrieval",
-      "task_number": 9
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "mrr",
@@ -434,32 +322,6 @@
       "task_label": "Cross-Modal Retrieval",
       "task_number": 9
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "r2",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "modality_reconstruction",
-      "task_label": "Cross-Modal Reconstruction",
-      "task_number": 10
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "r2",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "modality_reconstruction",
-      "task_label": "Cross-Modal Reconstruction",
-      "task_number": 10
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "r2",
@@ -525,32 +387,6 @@
       "task_label": "Temporal Order Verification",
       "task_number": 11
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "f1",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "misalignment_detection",
-      "task_label": "Multimodal Synchronization Detection",
-      "task_number": 12
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "f1",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "misalignment_detection",
-      "task_label": "Multimodal Synchronization Detection",
-      "task_number": 12
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "f1",
@@ -577,19 +413,6 @@
       "task_label": "Multimodal Synchronization Detection",
       "task_number": 12
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 887 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "long_horizon_next_action",
-      "task_label": "Long-Horizon Next-Action Forecasting",
-      "task_number": 13
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "macro_f1",
@@ -616,19 +439,6 @@
       "task_label": "Long-Horizon Next-Action Forecasting",
       "task_number": 13
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 651 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "next_subtask_forecast",
-      "task_label": "Long-Horizon Next-Subtask Forecasting",
-      "task_number": 14
-    },
     {
       "method": "Cosmos3-Super Reasoner",
       "metric_key": "macro_f1",
@@ -656,11 +466,11 @@
       "task_number": 14
     },
     {
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "metric_key": "macro_f1",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
       "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
@@ -669,11 +479,11 @@
       "task_number": 15
     },
     {
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "metric_key": "macro_f1",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
       "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_neural_mlp",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
@@ -720,19 +530,6 @@
       "task_label": "Interaction Text Prediction",
       "task_number": 15
     },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "macro_f1",
-      "reason": "train class count 3058 exceeds --max-neural-classes 512",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "action_object_relation",
-      "task_label": "Action-Object Relation Prediction",
-      "task_number": 16
-    },
     {
       "method": "Cosmos3-Nano Future Window",
       "metric_key": "macro_f1",
@@ -772,32 +569,6 @@
       "task_label": "Future Object-Set Forecasting",
       "task_number": 17
     },
-    {
-      "method": "128ep Metadata Simple",
-      "metric_key": "mae",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-      "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "task_id": "imu_to_hand_pose",
-      "task_label": "IMU-to-Hand Pose Reconstruction",
-      "task_number": 18
-    },
-    {
-      "method": "128ep Metadata NN",
-      "metric_key": "mae",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-      "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
-      "series_id": "metadata128_neural_mlp",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "task_id": "imu_to_hand_pose",
-      "task_label": "IMU-to-Hand Pose Reconstruction",
-      "task_number": 18
-    },
     {
       "method": "Qwen3-Omni v6 LoRA",
       "metric_key": "mae",
@@ -838,11 +609,11 @@
       "task_number": 18
     },
     {
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "metric_key": "mrr",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
       "recommended_next_step": "Export the missing target field for this 128-episode method, then rerun the same train/validation/test split.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
@@ -851,11 +622,11 @@
       "task_number": 19
     },
     {
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "metric_key": "mrr",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
       "recommended_next_step": "Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score.",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "series_id": "metadata128_neural_mlp",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
@@ -975,8 +746,8 @@
     "method_count": 9,
     "method_task_record_count": 180,
     "proxy_scored_method_task_count": 4,
-    "scored_method_task_count": 127,
-    "scoreless_method_task_count": 53,
+    "scored_method_task_count": 143,
+    "scoreless_method_task_count": 37,
     "task_count": 20
   },
   "source_matrix": "docs/data/task_method_20_result_matrix.json",

metrics/task_method_20_result_matrix.json

@@ -1,11 +1,11 @@
 {
   "title": "Task Method 20-Result Matrix",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "task_count": 20,
   "method_count": 9,
   "method_task_record_count": 180,
-  "scored_method_task_count": 133,
+  "scored_method_task_count": 143,
   "series": [
     {
       "id": "minimal",
@@ -55,50 +55,50 @@
     },
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -264,7 +264,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -274,7 +274,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -282,7 +282,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -292,7 +292,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -426,7 +426,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -436,7 +436,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -444,7 +444,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -454,7 +454,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -588,7 +588,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -598,7 +598,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -606,7 +606,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -616,7 +616,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -750,7 +750,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -760,7 +760,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -768,7 +768,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -778,7 +778,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -912,36 +912,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -1074,7 +1074,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1084,7 +1084,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1092,7 +1092,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1102,7 +1102,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1236,7 +1236,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1246,7 +1246,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1254,7 +1254,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1264,7 +1264,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1398,7 +1398,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1408,7 +1408,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1416,7 +1416,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1426,7 +1426,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1560,36 +1560,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -1722,36 +1722,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -1884,7 +1884,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1894,7 +1894,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -1902,7 +1902,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -1912,7 +1912,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2046,36 +2046,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -2208,7 +2208,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2218,7 +2218,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2226,7 +2226,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2236,7 +2236,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2370,7 +2370,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2380,7 +2380,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2388,7 +2388,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2398,7 +2398,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2532,7 +2532,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -2542,7 +2542,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -2550,7 +2550,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -2560,8 +2560,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -2694,7 +2694,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2704,7 +2704,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2712,7 +2712,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2722,7 +2722,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2856,7 +2856,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2866,7 +2866,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2874,7 +2874,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2884,7 +2884,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3018,36 +3018,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -3180,7 +3180,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -3190,7 +3190,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -3198,7 +3198,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -3208,8 +3208,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -3342,7 +3342,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3352,7 +3352,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3360,7 +3360,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3370,7 +3370,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

metrics/task_surface_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:25+00:00",
+  "generated_at_utc": "2026-06-18T12:54:18+00:00",
   "summary": {
     "task_count": 12,
     "expected_task_count": 12,

metrics/unified_task_model_radar.json

@@ -1,18 +1,18 @@
 {
   "title": "Unified 20-Task Model Radar",
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:07:15+00:00",
+  "generated_at_utc": "2026-06-18T12:52:26+00:00",
   "task_count": 20,
   "method_count": 9,
   "method_task_record_count": 180,
-  "scored_method_task_count": 133,
+  "scored_method_task_count": 143,
   "normalization_policy": {
     "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
     "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
     "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
     "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
     "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-    "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+    "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
     "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export."
   },
   "series": [
@@ -64,50 +64,50 @@
     },
     {
       "id": "metadata128_simple",
-      "label": "128ep Metadata Simple",
+      "label": "128ep Aligned Simple",
       "short_label": "128-S",
       "color": "#ffd166",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "9 6",
-      "method_detail": "128-episode JSONL metadata/text simple baselines.",
+      "method_detail": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "scored": 13,
-        "unsupported_without_required_target": 7
+        "scored": 18,
+        "unsupported_without_required_target": 2
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
       "id": "metadata128_neural_mlp",
-      "label": "128ep Metadata NN",
+      "label": "128ep Aligned NN",
       "short_label": "128-NN",
       "color": "#f472b6",
-      "kind": "partial_128_episode_metadata_baseline",
-      "scope": "128 selected episodes, JSONL metadata/text only",
+      "kind": "partial_128_episode_aligned_baseline",
+      "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
       "stroke_dasharray": "3 6",
-      "method_detail": "128-episode JSONL metadata/text MLP baselines.",
+      "method_detail": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
       "plotted_as": "colored point overlay",
       "result_record_count": 20,
-      "scored_task_count": 13,
-      "covered_task_count": 13,
+      "scored_task_count": 18,
+      "covered_task_count": 18,
       "proxy_scored_task_count": 0,
-      "scoreless_task_count": 7,
-      "unsupported_task_count": 7,
+      "scoreless_task_count": 2,
+      "unsupported_task_count": 2,
       "not_evaluated_task_count": 0,
       "status_counts": {
-        "not_supported_by_metadata_only_package": 7,
-        "scored": 13
+        "not_supported_by_metadata_only_package": 2,
+        "scored": 18
       },
-      "coverage_fraction": 0.65,
+      "coverage_fraction": 0.9,
       "result_record_fraction": 1.0
     },
     {
@@ -301,7 +301,7 @@
           "raw": 0.008252821966746326,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008252821966746326,
@@ -312,7 +312,7 @@
           "raw": 0.004175793689174209,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004175793689174209,
@@ -401,7 +401,7 @@
           "raw": 0.00019512195121951218,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.00019512195121951218,
@@ -412,7 +412,7 @@
           "raw": 7.207207207207208e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 7.207207207207208e-05,
@@ -523,7 +523,7 @@
           "raw": 0.29652162550029315,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.29652162550029315,
@@ -534,7 +534,7 @@
           "raw": 0.4841733292368365,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4841733292368365,
@@ -634,7 +634,7 @@
           "raw": 0.006514774539765508,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.006514774539765508,
@@ -645,7 +645,7 @@
           "raw": 0.004910507980164745,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004910507980164745,
@@ -709,15 +709,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 8.817333221435547,
           "metric_key": "mpjpe",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.012231610603598841,
+          "raw_text": "8.817",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.429434210062027,
+          "metric_key": "mpjpe",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.25114484128127007,
+          "raw_text": "0.4294",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.2729249894618988,
@@ -741,17 +752,6 @@
           "raw_text": "0.1848",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mpjpe",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "mpjpe",
@@ -856,7 +856,7 @@
           "raw": 0.4381481308057444,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4381481308057444,
@@ -867,7 +867,7 @@
           "raw": 0.5682695682695682,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.5682695682695682,
@@ -956,7 +956,7 @@
           "raw": 0.17764578833693304,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17764578833693304,
@@ -967,7 +967,7 @@
           "raw": 0.18662723837686876,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.18662723837686876,
@@ -1056,7 +1056,7 @@
           "raw": 0.002332374220713973,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.002332374220713973,
@@ -1067,7 +1067,7 @@
           "raw": 0.008236799389123917,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.008236799389123917,
@@ -1175,15 +1175,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.002587692579254508,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.002587692579254508,
+          "raw_text": "0.0026",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.0026067993603646755,
+          "metric_key": "mrr",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0026067993603646755,
+          "raw_text": "0.0026",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.003459817497059703,
@@ -1207,17 +1218,6 @@
           "raw_text": "0.0025",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mrr",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "cosmos3_super_reasoner": {
           "raw": null,
           "metric_key": "mrr",
@@ -1264,15 +1264,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": -190.66106203944798,
           "metric_key": "r2",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-190.66",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": -0.43481132003942147,
+          "metric_key": "r2",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.0,
+          "raw_text": "-0.4348",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": -1.3450960391924882,
@@ -1296,17 +1307,6 @@
           "raw_text": "-1.397",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "r2",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "r2",
@@ -1389,7 +1389,7 @@
           "raw": 0.4198864140782312,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.4198864140782312,
@@ -1400,7 +1400,7 @@
           "raw": 0.8252408266656923,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.8252408266656923,
@@ -1497,15 +1497,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.49980060227663614,
           "metric_key": "f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.49980060227663614,
+          "raw_text": "0.4998",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.7773773780941162,
+          "metric_key": "f1",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.7773773780941162,
+          "raw_text": "0.7774",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.4958867673901769,
@@ -1529,17 +1540,6 @@
           "raw_text": "0.8273",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "f1",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "cosmos3_super_reasoner": {
           "raw": null,
           "metric_key": "f1",
@@ -1611,7 +1611,7 @@
           "raw": 0.004579592783699693,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.004579592783699693,
@@ -1622,7 +1622,7 @@
           "raw": 0.0029821307969142615,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0029821307969142615,
@@ -1722,7 +1722,7 @@
           "raw": 0.0001206030150753769,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0001206030150753769,
@@ -1733,7 +1733,7 @@
           "raw": 2.086049543676662e-05,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 2.086049543676662e-05,
@@ -1822,7 +1822,7 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
           "normalized_score": null,
@@ -1855,9 +1855,9 @@
           "raw": null,
           "metric_key": "macro_f1",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -1955,7 +1955,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -1966,7 +1966,7 @@
           "raw": 0.0,
           "metric_key": "macro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.0,
@@ -2055,7 +2055,7 @@
           "raw": 0.17656983343047333,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17656983343047333,
@@ -2066,7 +2066,7 @@
           "raw": 0.17418550827844048,
           "metric_key": "micro_f1",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.17418550827844048,
@@ -2152,15 +2152,26 @@
           "status_label": "scored"
         },
         "metadata128_simple": {
-          "raw": null,
+          "raw": 0.2294670194387436,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "unsupported_without_required_target",
-          "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "unsupported"
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.18324815505876868,
+          "raw_text": "0.2295",
+          "status_label": "scored"
+        },
+        "metadata128_neural_mlp": {
+          "raw": 0.2555866539478302,
+          "metric_key": "mae",
+          "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+          "scope": "multi_episode_128_aligned_sensor_block_baseline",
+          "status": "scored",
+          "reason": null,
+          "normalized_score": 0.16452114110609004,
+          "raw_text": "0.2556",
+          "status_label": "scored"
         },
         "raw128_simple": {
           "raw": 0.22941437363624573,
@@ -2184,17 +2195,6 @@
           "raw_text": "0.2530",
           "status_label": "scored"
         },
-        "metadata128_neural_mlp": {
-          "raw": null,
-          "metric_key": "mae",
-          "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
-          "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
-          "normalized_score": null,
-          "raw_text": "n/a",
-          "status_label": "not supported"
-        },
         "qwen3_omni_v6_lora": {
           "raw": null,
           "metric_key": "mae",
@@ -2266,7 +2266,7 @@
           "raw": null,
           "metric_key": "mrr",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "unsupported_without_required_target",
           "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
           "normalized_score": null,
@@ -2299,9 +2299,9 @@
           "raw": null,
           "metric_key": "mrr",
           "source": null,
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "not_supported_by_metadata_only_package",
-          "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
+          "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required",
           "normalized_score": null,
           "raw_text": "n/a",
           "status_label": "not supported"
@@ -2388,7 +2388,7 @@
           "raw": 624.8108520507812,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.016864874132806403,
@@ -2399,7 +2399,7 @@
           "raw": 41.4664421081543,
           "metric_key": "mae",
           "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-          "scope": "multi_episode_128_metadata_baseline",
+          "scope": "multi_episode_128_aligned_baseline",
           "status": "scored",
           "reason": null,
           "normalized_score": 0.25411768748242325,
@@ -2456,18 +2456,18 @@
   "model_branch_cards": [
     {
       "id": "metadata128_simple",
-      "title": "128ep Metadata Simple",
+      "title": "128ep Aligned Simple",
       "status": "a100_rerun_pass",
-      "coverage": "20 records / 13 scored JSONL-supported axes",
+      "coverage": "20 records / 18 scored aligned axes",
       "headline": "34,269 rows; train/val/test 25,629/4,608/4,032",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/summary_report.json"
     },
     {
       "id": "metadata128_neural_mlp",
-      "title": "128ep Metadata NN",
+      "title": "128ep Aligned NN",
       "status": "a100_rerun_pass",
-      "coverage": "20 records / 13 scored JSONL-supported axes",
-      "headline": "compact MLP heads over metadata/text features",
+      "coverage": "20 records / 18 scored aligned axes",
+      "headline": "compact MLP heads over metadata/text and staged block features",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/summary_report.json"
     },
     {
@@ -2562,7 +2562,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2572,7 +2572,7 @@
       "normalized_score": 0.008252821966746326,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2580,7 +2580,7 @@
       "task_id": "timeline_action",
       "task_label": "Action Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2590,7 +2590,7 @@
       "normalized_score": 0.004175793689174209,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2724,7 +2724,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2734,7 +2734,7 @@
       "normalized_score": 0.00019512195121951218,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2742,7 +2742,7 @@
       "task_id": "timeline_subtask",
       "task_label": "Procedure Step Recognition",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2752,7 +2752,7 @@
       "normalized_score": 7.207207207207208e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/timeline_subtask/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2886,7 +2886,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2896,7 +2896,7 @@
       "normalized_score": 0.29652162550029315,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -2904,7 +2904,7 @@
       "task_id": "transition_detection",
       "task_label": "Action Boundary Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -2914,7 +2914,7 @@
       "normalized_score": 0.4841733292368365,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/transition_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3048,7 +3048,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3058,7 +3058,7 @@
       "normalized_score": 0.006514774539765508,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3066,7 +3066,7 @@
       "task_id": "next_action",
       "task_label": "Next-Action Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3076,7 +3076,7 @@
       "normalized_score": 0.004910507980164745,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3210,36 +3210,36 @@
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 8.817333221435547,
+      "raw_text": "8.817",
+      "normalized_score": 0.012231610603598841,
       "metric_key": "mpjpe",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/hand_trajectory_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires future hand-joint trajectories from raw sensor feature NPZ blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
       "task_id": "hand_trajectory_forecast",
       "task_label": "Hand Trajectory Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.429434210062027,
+      "raw_text": "0.4294",
+      "normalized_score": 0.25114484128127007,
       "metric_key": "mpjpe",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/hand_trajectory_forecast/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 5,
@@ -3372,7 +3372,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3382,7 +3382,7 @@
       "normalized_score": 0.4381481308057444,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3390,7 +3390,7 @@
       "task_id": "contact_prediction",
       "task_label": "Contact State Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3400,7 +3400,7 @@
       "normalized_score": 0.5682695682695682,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/contact_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3534,7 +3534,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3544,7 +3544,7 @@
       "normalized_score": 0.17764578833693304,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3552,7 +3552,7 @@
       "task_id": "object_relevance",
       "task_label": "Object Relevance Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3562,7 +3562,7 @@
       "normalized_score": 0.18662723837686876,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3696,7 +3696,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3706,7 +3706,7 @@
       "normalized_score": 0.002332374220713973,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3714,7 +3714,7 @@
       "task_id": "caption_grounding",
       "task_label": "Language Grounding",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -3724,7 +3724,7 @@
       "normalized_score": 0.008236799389123917,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/caption_grounding/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -3858,36 +3858,36 @@
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.002587692579254508,
+      "raw_text": "0.0026",
+      "normalized_score": 0.002587692579254508,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/cross_modal_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires paired motion/IMU/camera/audio/depth feature blocks, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
       "task_id": "cross_modal_retrieval",
       "task_label": "Cross-Modal Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.0026067993603646755,
+      "raw_text": "0.0026",
+      "normalized_score": 0.0026067993603646755,
       "metric_key": "mrr",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/cross_modal_retrieval/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 9,
@@ -4020,36 +4020,36 @@
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -190.66106203944798,
+      "raw_text": "-190.66",
+      "normalized_score": 0.0,
       "metric_key": "r2",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/modality_reconstruction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires source and target modality feature blocks such as depth/video vectors, which are not in the public 128 package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
       "task_id": "modality_reconstruction",
       "task_label": "Cross-Modal Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": -0.43481132003942147,
+      "raw_text": "-0.4348",
+      "normalized_score": 0.0,
       "metric_key": "r2",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/modality_reconstruction/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 10,
@@ -4182,7 +4182,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4192,7 +4192,7 @@
       "normalized_score": 0.4198864140782312,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4200,7 +4200,7 @@
       "task_id": "temporal_order",
       "task_label": "Temporal Order Verification",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4210,7 +4210,7 @@
       "normalized_score": 0.8252408266656923,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/temporal_order/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4344,36 +4344,36 @@
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.49980060227663614,
+      "raw_text": "0.4998",
+      "normalized_score": 0.49980060227663614,
       "metric_key": "f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/misalignment_detection/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
       "task_id": "misalignment_detection",
       "task_label": "Multimodal Synchronization Detection",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.7773773780941162,
+      "raw_text": "0.7774",
+      "normalized_score": 0.7773773780941162,
       "metric_key": "f1",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/misalignment_detection/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 12,
@@ -4506,7 +4506,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4516,7 +4516,7 @@
       "normalized_score": 0.004579592783699693,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4524,7 +4524,7 @@
       "task_id": "long_horizon_next_action",
       "task_label": "Long-Horizon Next-Action Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4534,7 +4534,7 @@
       "normalized_score": 0.0029821307969142615,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4668,7 +4668,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4678,7 +4678,7 @@
       "normalized_score": 0.0001206030150753769,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4686,7 +4686,7 @@
       "task_id": "next_subtask_forecast",
       "task_label": "Long-Horizon Next-Subtask Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -4696,7 +4696,7 @@
       "normalized_score": 2.086049543676662e-05,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -4830,7 +4830,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -4840,7 +4840,7 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
     },
     {
@@ -4848,7 +4848,7 @@
       "task_id": "interaction_text_prediction",
       "task_label": "Interaction Text Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -4858,8 +4858,8 @@
       "normalized_score": null,
       "metric_key": "macro_f1",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 15,
@@ -4992,7 +4992,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5002,7 +5002,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5010,7 +5010,7 @@
       "task_id": "action_object_relation",
       "task_label": "Action-Object Relation Prediction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5020,7 +5020,7 @@
       "normalized_score": 0.0,
       "metric_key": "macro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5154,7 +5154,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5164,7 +5164,7 @@
       "normalized_score": 0.17656983343047333,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5172,7 +5172,7 @@
       "task_id": "object_set_forecast",
       "task_label": "Future Object-Set Forecasting",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5182,7 +5182,7 @@
       "normalized_score": 0.17418550827844048,
       "metric_key": "micro_f1",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5316,36 +5316,36 @@
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
-      "status": "unsupported_without_required_target",
-      "status_label": "unsupported",
-      "scored": false,
+      "method": "128ep Aligned Simple",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2294670194387436,
+      "raw_text": "0.2295",
+      "normalized_score": 0.18324815505876868,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
       "task_id": "imu_to_hand_pose",
       "task_label": "IMU-to-Hand Pose Reconstruction",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
-      "status": "not_supported_by_metadata_only_package",
-      "status_label": "not supported",
-      "scored": false,
+      "method": "128ep Aligned NN",
+      "status": "scored",
+      "status_label": "scored",
+      "scored": true,
       "proxy_scored": false,
-      "raw": null,
-      "raw_text": "n/a",
-      "normalized_score": null,
+      "raw": 0.2555866539478302,
+      "raw_text": "0.2556",
+      "normalized_score": 0.16452114110609004,
       "metric_key": "mae",
-      "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/imu_to_hand_pose/metrics.json",
+      "scope": "multi_episode_128_aligned_sensor_block_baseline",
+      "reason": null
     },
     {
       "task_number": 18,
@@ -5478,7 +5478,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "unsupported_without_required_target",
       "status_label": "unsupported",
       "scored": false,
@@ -5488,7 +5488,7 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
     },
     {
@@ -5496,7 +5496,7 @@
       "task_id": "camera_view_sync_retrieval",
       "task_label": "Camera-View Synchronization Retrieval",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "not_supported_by_metadata_only_package",
       "status_label": "not supported",
       "scored": false,
@@ -5506,8 +5506,8 @@
       "normalized_score": null,
       "metric_key": "mrr",
       "source": null,
-      "scope": "multi_episode_128_metadata_baseline",
-      "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
+      "scope": "multi_episode_128_aligned_baseline",
+      "reason": "the 128-episode aligned rerun did not produce this task target; raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
     },
     {
       "task_number": 19,
@@ -5640,7 +5640,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_simple",
-      "method": "128ep Metadata Simple",
+      "method": "128ep Aligned Simple",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5650,7 +5650,7 @@
       "normalized_score": 0.016864874132806403,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {
@@ -5658,7 +5658,7 @@
       "task_id": "time_to_transition",
       "task_label": "Time-to-Next-Transition Regression",
       "series_id": "metadata128_neural_mlp",
-      "method": "128ep Metadata NN",
+      "method": "128ep Aligned NN",
       "status": "scored",
       "status_label": "scored",
       "scored": true,
@@ -5668,7 +5668,7 @@
       "normalized_score": 0.25411768748242325,
       "metric_key": "mae",
       "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
-      "scope": "multi_episode_128_metadata_baseline",
+      "scope": "multi_episode_128_aligned_baseline",
       "reason": null
     },
     {

metrics/website_integrity.json

@@ -1,6 +1,6 @@
 {
   "status": "pass",
-  "generated_at_utc": "2026-06-18T12:09:46+00:00",
+  "generated_at_utc": "2026-06-18T12:54:19+00:00",
   "docs_root": "docs",
   "site_base": "/ropedia-xperience-10m-task-suite/",
   "summary": {
@@ -301,7 +301,7 @@
     },
     {
       "path": "data/artifact_index.json",
-      "bytes": 116110,
+      "bytes": 116111,
       "top_level_type": "dict"
     },
     {
@@ -316,7 +316,7 @@
     },
     {
       "path": "data/episode128_task_model_radar.json",
-      "bytes": 186443,
+      "bytes": 185447,
       "top_level_type": "dict"
     },
     {
@@ -351,7 +351,7 @@
     },
     {
       "path": "data/mirror_parity.json",
-      "bytes": 994053,
+      "bytes": 1059014,
       "top_level_type": "dict"
     },
     {
@@ -471,7 +471,7 @@
     },
     {
       "path": "data/single_episode_task_model_radar.json",
-      "bytes": 50973,
+      "bytes": 51064,
       "top_level_type": "dict"
     },
     {
@@ -486,12 +486,12 @@
     },
     {
       "path": "data/task_method_20_gap_audit.json",
-      "bytes": 46902,
+      "bytes": 35883,
       "top_level_type": "dict"
     },
     {
       "path": "data/task_method_20_result_matrix.json",
-      "bytes": 129242,
+      "bytes": 128794,
       "top_level_type": "dict"
     },
     {
@@ -526,7 +526,7 @@
     },
     {
       "path": "data/unified_task_model_radar.json",
-      "bytes": 230297,
+      "bytes": 229299,
       "top_level_type": "dict"
     },
     {
@@ -571,7 +571,7 @@
     {
       "path": "assets/charts/episode128_task_model_radar.svg",
       "exists": true,
-      "bytes": 45937,
+      "bytes": 47540,
       "format": "SVG",
       "has_viewbox": true
     },
@@ -641,7 +641,7 @@
     {
       "path": "assets/charts/unified_task_model_radar.svg",
       "exists": true,
-      "bytes": 51953,
+      "bytes": 53553,
       "format": "SVG",
       "has_viewbox": true
     },
@@ -752,7 +752,7 @@
     {
       "path": "assets/task_suite_infographic.png",
       "exists": true,
-      "bytes": 2627286,
+      "bytes": 1591194,
       "width": 1800,
       "height": 6600,
       "format": "PNG"

scripts/build_unified_task_model_radar.py

@@ -114,19 +114,19 @@ SERIES = {
         "stroke_dasharray": None,
     },
     "metadata128_simple": {
-        "label": "128ep Metadata Simple",
+        "label": "128ep Aligned Simple",
         "short_label": "128-S",
         "color": "#ffd166",
-        "kind": "partial_128_episode_metadata_baseline",
-        "scope": "128 selected episodes, JSONL metadata/text only",
+        "kind": "partial_128_episode_aligned_baseline",
+        "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
         "stroke_dasharray": "9 6",
     },
     "metadata128_neural_mlp": {
-        "label": "128ep Metadata NN",
+        "label": "128ep Aligned NN",
         "short_label": "128-NN",
         "color": "#f472b6",
-        "kind": "partial_128_episode_metadata_baseline",
-        "scope": "128 selected episodes, JSONL metadata/text only",
+        "kind": "partial_128_episode_aligned_baseline",
+        "scope": "128 selected episodes, JSONL metadata/text plus staged sensor-block targets where available",
         "stroke_dasharray": "3 6",
     },
     "raw128_simple": {
@@ -254,8 +254,8 @@ SHORT_TASK_LABELS = {
 METHOD_DETAILS = {
     "minimal": "Single-episode simple heads over the public sample split.",
     "neural_mlp": "Single-episode compact PyTorch MLP heads on the same 20 task contracts.",
-    "metadata128_simple": "128-episode JSONL metadata/text simple baselines.",
-    "metadata128_neural_mlp": "128-episode JSONL metadata/text MLP baselines.",
+    "metadata128_simple": "128-episode aligned simple baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
+    "metadata128_neural_mlp": "128-episode aligned MLP baselines: JSONL metadata/text tasks plus staged sensor-block tasks where the processed target exists.",
     "raw128_simple": "128-episode 4430-dim sensor NPZ simple heads; tasks 15/19 use compact proxies.",
     "raw128_neural_mlp": "128-episode 4430-dim sensor NPZ MLP heads; tasks 15/19 use compact proxies.",
     "qwen3_omni_v6_lora": "Verified held-out Qwen3-Omni v6 LoRA metrics, plus task 16 and any completed private-GPU future-task probes scored from task-specific JSON.",
@@ -322,12 +322,12 @@ def read_a100_metadata_record(task_id: str, *, neural: bool = False) -> dict[str
         "raw": score,
         "metric_key": payload.get("primary_metric"),
         "source": str(path.relative_to(ROOT)),
-        "scope": "multi_episode_128_metadata_baseline",
+        "scope": payload.get("scope") or "multi_episode_128_aligned_baseline",
         "status": "scored" if status == "pass" and score is not None else "unsupported_without_required_target",
         "reason": payload.get("reason")
         or payload.get("error")
         or (
-            "metadata-only package has a metrics artifact for this task, but it does not contain a numeric public score"
+            "the 128-episode aligned artifact for this task does not contain a numeric public score"
             if status != "pass"
             else None
         ),
@@ -398,10 +398,10 @@ def make_missing_record(series_id: str, task_id: str, metric_key: str | None) ->
     if series_id.startswith("metadata128"):
         status = "not_supported_by_metadata_only_package"
         reason = (
-            "the 128-episode metadata/text rerun did not produce this task target; "
-            "raw sensor blocks or a task-specific metadata target builder are required"
+            "the 128-episode aligned rerun did not produce this task target; "
+            "raw interaction text, paired camera-view embeddings, or a task-specific target builder is required"
         )
-        scope = "multi_episode_128_metadata_baseline"
+        scope = "multi_episode_128_aligned_baseline"
     elif series_id in {"qwen3_omni_v6_lora", "cosmos3_super_reasoner", "cosmos3_nano_future_window"}:
         status = "not_evaluated_in_verified_package"
         reason = (
@@ -745,7 +745,7 @@ def build_payload() -> dict[str, Any]:
             "raw_values": "raw metric values, metric keys, and sources are retained in this JSON; the SVG is an overview, not a replacement for the metric table",
             "result_record_policy": "every method has 20 task records; records without a numeric score carry explicit unsupported/not-evaluated status and reason fields",
             "foundation_model_overlay": "Qwen3/Cosmos points are plotted only on task-aligned axes. Scoreless records mean the public result does not evaluate that task contract.",
-            "metadata_128_overlay": "128-episode metadata baselines have 20 records, but numeric scores only where the public JSONL contains enough task labels without raw feature blocks.",
+            "metadata_128_overlay": "128-episode aligned baselines have 20 records. Numeric scores come from JSONL metadata/text tasks plus staged sensor-block targets when the processed target exists; raw interaction text and paired camera-view embeddings remain explicit gaps.",
             "raw_128_overlay": "128-episode raw-feature baselines use staged sensor NPZ features. Eighteen axes use direct task targets; interaction text and camera-view sync are completed with documented compact proxies because raw interaction strings and paired video-view embeddings are absent from the 128 export.",
         },
         "series": series_records,
@@ -753,18 +753,18 @@ def build_payload() -> dict[str, Any]:
         "model_branch_cards": [
             {
                 "id": "metadata128_simple",
-                "title": "128ep Metadata Simple",
+                "title": "128ep Aligned Simple",
                 "status": "a100_rerun_pass",
-                "coverage": f"20 records / {next(item for item in series_records if item['id'] == 'metadata128_simple')['scored_task_count']} scored JSONL-supported axes",
+                "coverage": f"20 records / {next(item for item in series_records if item['id'] == 'metadata128_simple')['scored_task_count']} scored aligned axes",
                 "headline": "34,269 rows; train/val/test 25,629/4,608/4,032",
                 "source": str((METADATA128_BASELINE_DIR / "summary_report.json").relative_to(ROOT)),
             },
             {
                 "id": "metadata128_neural_mlp",
-                "title": "128ep Metadata NN",
+                "title": "128ep Aligned NN",
                 "status": "a100_rerun_pass",
-                "coverage": f"20 records / {next(item for item in series_records if item['id'] == 'metadata128_neural_mlp')['scored_task_count']} scored JSONL-supported axes",
-                "headline": "compact MLP heads over metadata/text features",
+                "coverage": f"20 records / {next(item for item in series_records if item['id'] == 'metadata128_neural_mlp')['scored_task_count']} scored aligned axes",
+                "headline": "compact MLP heads over metadata/text and staged block features",
                 "source": str((METADATA128_BASELINE_DIR / "summary_report.json").relative_to(ROOT)),
             },
             {

scripts/omni/run_128_task_baselines.py

@@ -1463,12 +1463,28 @@ def unsupported_record(task_id: str, out_root: Path, reason: str, primary_metric
 
 
 def build_markdown(summary: dict[str, Any]) -> str:
+    sensor_completion = bool((summary.get("feature_contract") or {}).get("sensor_block_completion"))
+    source_sentence = (
+        "The aligned runner uses the derived Qwen JSONL export for metadata/text tasks and staged processed sensor NPZ blocks only for the explicitly listed block-completion tasks. It still does not use raw Xperience-10M videos, raw annotation HDF5 files, Qwen weights, or LoRA weights."
+        if sensor_completion
+        else "The runner uses the derived Qwen JSONL export and public-safe metadata. It does not use raw Xperience-10M videos, HDF5 files, sensor NPZ blocks, Qwen weights, or LoRA weights."
+    )
+    unsupported_sentence = (
+        "Tasks still marked unsupported require raw annotation interaction text or paired camera-view embeddings that are absent from the staged 128 export."
+        if sensor_completion
+        else "Tasks marked `unsupported_without_raw_128_feature_blocks` still need the 128-run sensor feature NPZ blocks to reproduce the single-episode feature-level target exactly."
+    )
+    interpretation_sentence = (
+        "The trainable scores combine JSONL metadata/text tasks with staged sensor-block completion tasks. They are useful for checking split alignment, label difficulty, train/test target coverage, and whether the Qwen diagnostic run is being compared against the same 96/16/16 episode setup."
+        if sensor_completion
+        else "The trainable scores are metadata/text baselines, not replacements for full raw-modality baselines. They are useful for checking split alignment, label difficulty, train/test label coverage, and whether the Qwen diagnostic run is being compared against the same 96/16/16 episode setup."
+    )
     lines = [
         "# 128-Episode Aligned Baselines",
         "",
         "These results align the earlier simple and neural baseline framing to the same selected 128-episode split used by the Qwen3-Omni pilot.",
         "",
-        "The runner uses the derived Qwen JSONL export and public-safe metadata. It does not use raw Xperience-10M videos, HDF5 files, sensor NPZ blocks, Qwen weights, or LoRA weights.",
+        source_sentence,
         "",
         "## Split",
         "",
@@ -1502,9 +1518,9 @@ def build_markdown(summary: dict[str, Any]) -> str:
             "",
             "## Interpretation",
             "",
-            "The trainable scores are metadata/text baselines, not replacements for full raw-modality baselines. They are useful for checking split alignment, label difficulty, train/test label coverage, and whether the Qwen diagnostic run is being compared against the same 96/16/16 episode setup.",
+            interpretation_sentence,
             "",
-            "Tasks marked `unsupported_without_raw_128_feature_blocks` still need the 128-run sensor feature NPZ blocks to reproduce the single-episode feature-level target exactly.",
+            unsupported_sentence,
         ]
     )
     return "\n".join(lines) + "\n"