“Not Aligned” is Not “Malicious”: Being Careful about Hallucinations of Large Language Models’ Jailb

Last updated