AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models

arXiv:2310.04451v2 Announce Type: replace
Abstract: 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) …

abstract alignment arxiv attacks cs.ai cs.cl decision feedback however human human feedback jailbreak language language models language understanding large language large language models large models llms making prompts through tools type understanding