LLM Defenses Are Not Robust To Multi-turn Human Jailbreaks Yet
Li Nathaniel, Han Ziwen, Steneker Ian, Primack Willow, Goodside Riley, Zhang Hugh, Wang Zifan, Menghini Cristina, Yue Summer. Arxiv 2024
[Paper]
Recent large language model (LLM) defenses have greatly improved models’ ability to refuse harmful queries, even when adversarially attacked. However, LLM defenses are primarily evaluated against automated adversarial attacks in a single turn of conversation, an insufficient threat model for real-world malicious use. We demonstrate that multi-turn human jailbreaks uncover significant vulnerabilities, exceeding 70% attack success rate (ASR) on HarmBench against defenses that report single-digit ASRs with automated single-turn attacks. Human jailbreaks also reveal vulnerabilities in machine unlearning defenses, successfully recovering dual-use biosecurity knowledge from unlearned models. We compile these results into Multi-Turn Human Jailbreaks (MHJ), a dataset of 2,912 prompts across 537 multi-turn jailbreaks. We publicly release MHJ alongside a compendium of jailbreak tactics developed across dozens of commercial red teaming engagements, supporting research towards stronger LLM defenses.
Similar Work