Best-of-N Jail an algorithm that jailbreaks LLM across modalities

8 points| amram_art | 1 year ago |jplhughes.github.io