Distractor-Based Jailbreaking Attacks in Language Models and Associated Changes in Chain-of-Thought Content (Student Abstract).

Tate Rowney, Xuning Ying

Browse the full AAAI paper archive.