Exploiting the Index Gradients for Optimization-Based Jailbreaking on Large Language Models.

Jiahui Li, Yongchang Hao, Haoyu Xu, Xing Wang, Yu Hong

Browse the full COLING paper archive.