Extraltodeus/J-Wash

Jacobian-Brainwash : A manual alignment tool for large language models built on Anthropic's Jacobian Lens. Results are exportable.

⭐ 224 ⑂ 22 Python Apache-2.0 · 21 天前推送

abliteration ai anthropic artificial-intelligence huggingface large-language-models llm neural-networks pytorch research transformers visualization interpretability j-wash jacobian-lens mechanistic-interpretability model-editing steering

224

Watchers

0

贡献者

0

Commits

0

Releases

3

Open Issues

21 天前

最近推送

原文中文

暂无 README