I created a super harmful model ! :D (by tweaking it's J-Space!!!)

Soooo! Since Anthropic share their Jacobian-Lens a few days ago I went on and made a tool based on it which adds the possibilité to export a model which will have the same behavior after tweaking it's J-Space. This means manually alter the behavior and abliterate by using a human brain. I'm still working on it but couldn't wait to produce something first. SO After finally getting a working codebase I immediatly jumped and tried to make pretty pervy model PURELY in the name of science. Let me introduce you t

添加评论
点赞收藏
点踩分享查看原文
评论
?
参与讨论