Comment on Refusal in LLMs is mediated by a single direction
toxuin@lemmy.ca 1 year agoIt works in reverse too. You can make any LLM “forget” that it is even able to refuse anything.
Comment on Refusal in LLMs is mediated by a single direction
toxuin@lemmy.ca 1 year agoIt works in reverse too. You can make any LLM “forget” that it is even able to refuse anything.
Enkers@sh.itjust.works 1 year ago
Oh for sure, and that was the main point, but I just find LLMs that refuse to do anything at all hilarious.
I wonder how much work it’d be to use this to jailbreak llama3. I only started playing with local LLMs recently.