Can I read AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models on EtoBox?

AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models by Liu, Xiaogeng; Xu, Nan; Chen, Muhao; Xiao, Chaowei is a scholarly article available to read on EtoBox.

What is AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models about?

The aligned Large Language Models (LLMs) are powerful language understanding and decision-making tools that are created through extensive alignment with human feedback. However, these large models remain susceptible to jailbreak attacks, where adversaries manipulate prompts to elicit malicious outputs that should not be given by aligned LLMs. Investigating jailbreak prompts can lead us to delve into the limitations of LLMs and further guide us to secure them. Unfortunately, existing jailbreak techniques suffer from either (1) scalability issues, where attacks heavily rely on manual crafting of prompts, or (2) stealthiness problems, as attacks depend on token-based algorithms to generate prompts that are often semantically meaningless, making them susceptible to detection through basic perplexity testing. In light of these challenges, we intend to answer this question: Can we develop an approach that can automatically generate stealthy jailbreak prompts? In this paper, we introduce AutoDAN, a novel jailbreak attack against aligned LLMs. AutoDAN can automatically generate stealthy jailbreak prompts by the carefully designed hierarchical genetic algorithm. Extensive evaluations demons

Author

Liu, Xiaogeng; Xu, Nan; Chen, Muhao; Xiao, Chaowei

Published

2023

Language

EN