FinalFormal2 comments on Using GPT-Eliezer against ChatGPT Jailbreaking

[FinalFormal2]

Wouldn’t it be hilarious if a variant of this was all it took to have exceptional AI safety