{
  "title": "Ropedia Xperience-10M Task Suite Reviewer Packet",
  "version": "2026-06-01",
  "scope_status": {
    "validated_data": "one public Xperience-10M sample episode",
    "aligned_frames": 5821,
    "sliding_windows": 1161,
    "current_feature_dimensions": 8378,
    "core_task_count": 12,
    "neural_head_count": 12,
    "direction_extension_probe_count": 4,
    "raw_xperience10m_data_in_repo": false,
    "audio_feature_status": "Audio is present in the sample MP4 streams and shown in the figures, but the current baseline feature vector does not include an extracted audio block.",
    "qwen3_omni_32_episode_claim": false,
    "qwen3_omni_status": "Smoke-only until at least 32 valid episodes are available and held-out episode evaluation finishes."
  },
  "review_path": [
    {
      "step": 1,
      "question": "What exactly is being claimed?",
      "primary_artifacts": [
        "EVIDENCE_CONTRACT.md",
        "ARTIFACT_GUIDE.md",
        "docs/data/evidence_contract.json",
        "docs/data/artifact_index.json",
        "docs/data/mirror_parity.json",
        "docs/data/publication_audit.json",
        "docs/data/scope_claims_audit.json",
        "docs/data/website_integrity.json"
      ],
      "readout": "Single-episode task engineering, metrics, visualizations, publication hygiene, local website integrity, mirror parity, and scope-claim guarding are claimed; cross-episode generalization and 32-episode Qwen3-Omni quality are not claimed."
    },
    {
      "step": 2,
      "question": "How can the public pipeline be reproduced?",
      "primary_artifacts": [
        "REPRODUCIBILITY.md",
        "docs/data/reproducibility_matrix.json",
        "notes/reproducibility_audit.md"
      ],
      "readout": "The public sample pipeline has explicit commands, expected outputs, and a prior exact-match audit over the committed metrics."
    },
    {
      "step": 3,
      "question": "What is inside one model input?",
      "primary_artifacts": [
        "results/episode_task_suite/windows.csv",
        "results/episode_task_suite/feature_manifest.json",
        "results/episode_task_suite/available_modalities.json",
        "docs/data/modality_atlas.json"
      ],
      "readout": "The current model input is an 8,378-dimensional aligned window vector with explicit feature-block boundaries, and the readable atlas shows each public-sample modality without raw data redistribution."
    },
    {
      "step": 4,
      "question": "Do the task metrics have committed evidence?",
      "primary_artifacts": [
        "results/episode_task_suite/summary_report.json",
        "results/episode_task_suite/neural_mlp/",
        "docs/data/summary_metrics.json"
      ],
      "readout": "Each of the 12 tasks has minimal-head metrics and a matching neural MLP result over the same window contracts."
    },
    {
      "step": 5,
      "question": "How should this scale beyond one episode?",
      "primary_artifacts": [
        "results/omni_finetune/DATA_BLOCKER_REPORT.md",
        "results/omni_finetune/A100_HF_RELAY_STATUS.md",
        "scripts/omni/discover_xperience10m_sources.py"
      ],
      "readout": "The next milestone is a 32-episode held-out-episode Qwen3-Omni LoRA pilot after gated Xperience-10M access is available."
    }
  ],
  "artifact_guide": "ARTIFACT_GUIDE.md",
  "artifact_index": "docs/data/artifact_index.json",
  "reproducibility_matrix": "docs/data/reproducibility_matrix.json",
  "public_surfaces": {
    "github_repo": "https://github.com/ChaoYue0307/ropedia-xperience-10m-task-suite",
    "github_pages": "https://chaoyue0307.github.io/ropedia-xperience-10m-task-suite/",
    "hf_space": "https://huggingface.co/spaces/cy0307/ropedia-xperience-10m-task-suite",
    "hf_static_app": "https://cy0307-ropedia-xperience-10m-task-suite.static.hf.space/",
    "hf_artifacts": "https://huggingface.co/datasets/cy0307/ropedia-xperience-10m-task-suite-artifacts",
    "hf_model_baselines": "https://huggingface.co/cy0307/ropedia-xperience-10m-task-baselines"
  },
  "do_not_infer": [
    "Do not infer cross-environment generalization from the single public sample episode.",
    "Do not treat the Qwen3-Omni smoke run as a 32-episode fine-tune.",
    "Do not treat feature-vector reconstruction as pixel-depth, mesh, NeRF, or Gaussian reconstruction.",
    "Do not assume raw Xperience-10M data is redistributed in this repo."
  ]
}