DeTAM: Defending LLMs Against Jailbreak Attacks via Targeted Attention Modification.

Yu Li, Han Jiang, Zhihua Wei

Browse the full ACL paper archive.