Any-to-Any
Transformers
Safetensors
gemma4
image-text-to-text
obliteratus
abliterated
uncensored
multimodal
bfloat16
refusal-reduction
Instructions to use KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS") model = AutoModelForMultimodalLM.from_pretrained("KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update_final_release_summary
Browse files
evals/evaluation_summary.json
CHANGED
|
@@ -1,11 +1,24 @@
|
|
| 1 |
{
|
| 2 |
-
"schema_version":
|
| 3 |
"model_id": "KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS",
|
| 4 |
"release_status": "rc1",
|
| 5 |
"release_dtype": "bfloat16",
|
| 6 |
"base_model": "google/gemma-4-E4B-it",
|
| 7 |
"base_revision": "fa62d88df2e6df5efa9d26ad6b3beaea2765f0cd",
|
| 8 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
"method_disclosure": "withheld",
|
| 10 |
"hub_verification": {
|
| 11 |
"weight_commit": "e22d6fff730361b2efc150fefeedd5c4f47f488d",
|
|
@@ -23,7 +36,7 @@
|
|
| 23 |
"generation_smoke_prompt": "What is 12 + 5? Answer with only the integer.",
|
| 24 |
"generation_smoke_parsed_content": "17"
|
| 25 |
},
|
| 26 |
-
"
|
| 27 |
"temperature": 0.0,
|
| 28 |
"top_p": 1.0,
|
| 29 |
"top_k": 0,
|
|
|
|
| 1 |
{
|
| 2 |
+
"schema_version": 3,
|
| 3 |
"model_id": "KridgeDookie/Gemma-4-E4B-IT-ABLITERATED-UNCENSORED-PHILADELPHIA-CLASS",
|
| 4 |
"release_status": "rc1",
|
| 5 |
"release_dtype": "bfloat16",
|
| 6 |
"base_model": "google/gemma-4-E4B-it",
|
| 7 |
"base_revision": "fa62d88df2e6df5efa9d26ad6b3beaea2765f0cd",
|
| 8 |
+
"internal_comparison_model": "private internal NF4-reconstructed dense BF16 baseline; not a public model and not published in this repository",
|
| 9 |
+
"matched_public_capability": {
|
| 10 |
+
"artifact": "evals/matched_public_capability_lm_eval_0_4_12.json",
|
| 11 |
+
"artifact_sha256": "0200b25a5760a97762cb0eff3bfae60418acb992d12c4dbfc639c00055f42a4f",
|
| 12 |
+
"scope": "full matched candidate versus untouched upstream control; not a leaderboard",
|
| 13 |
+
"upstream_revision": "fa62d88df2e6df5efa9d26ad6b3beaea2765f0cd",
|
| 14 |
+
"primary_delta_percentage_points": {
|
| 15 |
+
"mmlu": -0.3062,
|
| 16 |
+
"hellaswag_acc_norm": 0.0498,
|
| 17 |
+
"truthfulqa_mc2": -1.2642,
|
| 18 |
+
"gsm8k_strict": -0.0758,
|
| 19 |
+
"winogrande": 0.0789
|
| 20 |
+
}
|
| 21 |
+
},
|
| 22 |
"method_disclosure": "withheld",
|
| 23 |
"hub_verification": {
|
| 24 |
"weight_commit": "e22d6fff730361b2efc150fefeedd5c4f47f488d",
|
|
|
|
| 36 |
"generation_smoke_prompt": "What is 12 + 5? Answer with only the integer.",
|
| 37 |
"generation_smoke_parsed_content": "17"
|
| 38 |
},
|
| 39 |
+
"internal_generation_protocol": {
|
| 40 |
"temperature": 0.0,
|
| 41 |
"top_p": 1.0,
|
| 42 |
"top_k": 0,
|