Diversity Helps Jailbreak Large Language Models.

Weiliang Zhao, Daniel Ben-Levi, Wei Hao, Junfeng Yang, Chengzhi Mao

Browse the full NAACL paper archive.