Skip to content

Fully automatic censorship removal for language models

Add model card when uploading to Hugging Face

Philipp Emanuel Weidmann2ff8dcba6bc5
diff --git a/src/heretic/main.py b/src/heretic/main.py
index 2f3a6cd..63c8d65 100644
--- a/src/heretic/main.py
+++ b/src/heretic/main.py
@@ -19,11 +19,12 @@ from accelerate.utils import (
is_sdaa_available,
is_xpu_available,
)
+from huggingface_hub import ModelCard
from .config import Settings
from .evaluator import Evaluator
from .model import Model
-from .utils import print
+from .utils import get_readme_intro, print
def main():
@@ -303,8 +304,29 @@ def main():
private = visibility == "Private"
print("Uploading model...")
+
model.model.push_to_hub(repo_id, private=private, token=token)
model.tokenizer.push_to_hub(repo_id, private=private, token=token)
+
+ # If the model path doesn't exist locally, it can be assumed
+ # to be a model hosted on the Hugging Face Hub, in which case
+ # we can retrieve the model card.
+ if not Path(settings.model).exists():
+ card = ModelCard.load(settings.model)
+ card.data.tags.append("heretic")
+ card.data.tags.append("uncensored")
+ card.data.tags.append("decensored")
+ card.text = (
+ get_readme_intro(
+ settings,
+ study,
+ evaluator.base_refusals,
+ evaluator.bad_prompts,
+ )
+ + card.text
+ )
+ card.push_to_hub(repo_id, token=token)
+
print(f"Model uploaded to [bold]{repo_id}[/].")
case "Chat with the model":
diff --git a/src/heretic/utils.py b/src/heretic/utils.py
index eea9ec0..e66ab4f 100644
--- a/src/heretic/utils.py
+++ b/src/heretic/utils.py
@@ -4,6 +4,7 @@
import gc
from typing import TypeVar
+import optuna
import torch
from accelerate.utils import (
is_mlu_available,
@@ -14,7 +15,7 @@ from accelerate.utils import (
from datasets import load_dataset
from rich.console import Console
-from .config import DatasetSpecification
+from .config import DatasetSpecification, Settings
print = Console(highlight=False).print
@@ -44,3 +45,37 @@ def empty_cache():
torch.musa.empty_cache()
gc.collect()
+
+
+def get_readme_intro(
+ settings: Settings,
+ study: optuna.Study,
+ base_refusals: int,
+ bad_prompts: list[str],
+) -> str:
+ refusal_percentage = (
+ study.best_trial.user_attrs["refusals"] / len(bad_prompts) * 100
+ )
+ base_refusal_percentage = base_refusals / len(bad_prompts) * 100
+
+ return f"""# This is a decensored version of [{settings.model}](https://huggingface.co/{settings.model}), made using [Heretic](https://github.com/p-e-w/heretic)
+
+## Abliteration parameters
+
+| Parameter | Value |
+| :---------------------- | :--------------------------------------------: |
+| **max_weight** | {study.best_params["max_weight"]:.4f} |
+| **max_weight_position** | {study.best_params["max_weight_position"]:.4f} |
+| **min_weight** | {study.best_params["min_weight"]:.4f} |
+| **min_weight_distance** | {study.best_params["min_weight_distance"]:.4f} |
+
+## Performance
+
+| Metric | This model | Original model ([{settings.model}](https://huggingface.co/{settings.model})) |
+| :---------------- | :------------------------------------------------: | :--------------------------------------------------------------------------: |
+| **KL divergence** | {study.best_trial.user_attrs["kl_divergence"]:.4f} | 0 *(by definition)* |
+| **Refusals** | {refusal_percentage:.1f} % | {base_refusal_percentage:.1f} % |
+
+-----
+
+"""