You Can't Eat Your Cake and Have It Too: The Performance Degradation of LLMs with Jailbreak Defense

21 January 2025

Papers citing "You Can't Eat Your Cake and Have It Too: The Performance Degradation of LLMs with Jailbreak Defense"

1 / 1 papers shown

Title
Output Constraints as Attack Surface: Exploiting Structured Generation to Bypass LLM Safety Mechanisms Shuoming Zhang Jiacheng Zhao Ruiyuan Xu Xiaobing Feng Huimin Cui AAML 39 2 0 31 Mar 2025