CHASE: Contextual History for Adaptive and Simple Exploitation in Large Language Model Jailbreaking

We propose Contextual History for Adaptive and Simple Exploitation (CHASE), a novel multi-turn method for Large Language Model (LLM) jailbreaking. Rather than directly attack an LLM that may be difficult to jailbreak, CHASE first collects jailbroken histories from an easy-to-jailbreak LLM and then transfers them to the target LLM. Through this history transfer process, CHASE misleads the target LLM into thinking that it is responsible for producing the jailbroken histories and increases the chances of successful jailbreaking by prompting it to continue the conversation. Extensive evaluations on mainstream LLMs show that CHASE consistently achieves higher attack success rates and demands fewer computational resources compared to existing methods.

Authors

• Zhiqiang Hao Nanjing University
• Chuanyi Li Nanjing University
• Ye Fan China Mobile Communications Group Jiangsu Co., Ltd
• Jun Cai Nanjing University
• Xiao Fu
• Shangqi Wang China Mobile Communications Group Jiangsu Co., Ltd
• Hao Shen Hefei University of Technology
• Jiao Yin China Mobile Communications Group Jiangsu Co., Ltd
• Jidong Ge
• Bin Luo Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, Anhui Provincial Key Laboratory of Multimodal Cognitive Computation, School of Computer Science and Technology, Anhui University, China
• Vincent Ng University of Texas at Dallas

Keywords

No keywords are indexed for this paper.