AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs
Xiaogeng Liu, Peiran Li, G. Edward Suh, Yevgeniy Vorobeychik, Zhuoqing Mao, Somesh Jha, Patrick McDaniel, Huan Sun, Bo Li, Chaowei Xiao
Abstract
In this paper, we propose AutoDAN-Turbo, a black-box jailbreak method that can automatically discover as many jailbreak strategies as possible from scratch, without any human intervention or predefined scopes (e.g., specified candidate strategies), and use them for red-teaming. As a result, AutoDAN-Turbo can significantly outperform baseline methods, achieving a 74.3% higher average attack success rate on public benchmarks. Notably, AutoDAN-Turbo achieves an 88.5 attack success rate on GPT-4-1106-turbo. In addition, AutoDAN-Turbo is a unified framework that can incorporate existing human-designed jailbreak strategies in a plug-and-play manner. By integrating human-designed strategies, AutoDAN-Turbo can even achieve a higher attack success rate of 93.4 on GPT-4-1106-turbo.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
65 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher cites Training language models to follow instructions with human feedback
- Lifelong Safety Alignment for Language Models cites GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher
- Lifelong Safety Alignment for Language Models cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- Lifelong Safety Alignment for Language Models cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
- Tree of Attacks: Jailbreaking Black-Box LLMs Automatically cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- Tree of Attacks: Jailbreaking Black-Box LLMs Automatically cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs cites GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs cites Tree of Attacks: Jailbreaking Black-Box LLMs Automatically
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs cites Training language models to follow instructions with human feedback
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMs cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
- AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration cites Tree of Attacks: Jailbreaking Black-Box LLMs Automatically
- AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection Attacks cites Tree of Attacks: Jailbreaking Black-Box LLMs Automatically
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency cites GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency cites Tree of Attacks: Jailbreaking Black-Box LLMs Automatically
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency cites Training language models to follow instructions with human feedback
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency cites AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
- AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models cites GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher
- AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
- AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models cites Training language models to follow instructions with human feedback
Cited by top-tier papers65
- AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack IntegrationAndy Zhou, Kevin Wu, Francesco Pinto, Zhaorun Chen et al.NeurIPS 2025 · 46 citations
- Lifelong Safety Alignment for Language ModelsHaoyu Wang, Yifei Zhao, Zeyu Qin, Chao Du et al.NeurIPS 2025 · 18 citations
- MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection AttacksGeorgios Syros, Evan Rose, Brian Grinstead, Christoph Kerschbaumer et al.USENIX Security 2026 · 18 citations
- Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task ConcurrencyYukun Jiang, Mingjie Li, Michael Backes, Yang ZhangNeurIPS 2025 · 17 citations
- GASP: Efficient Black-Box Generation of Adversarial Suffixes for Jailbreaking LLMsAdvik Raj Basani, Xiao ZhangNeurIPS 2025 · 16 citations
Builds on10
- Training language models to follow instructions with human feedbackLong Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida et al.NeurIPS 2022 · 24,707 citations
- HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust RefusalMantas Mazeika, Long Phan, Xuwang Yin, Andy Zou et al.ICML 2024 · 1,031 citations
- Tree of Attacks: Jailbreaking Black-Box LLMs AutomaticallyAnay Mehrotra, Manolis Zampetakis, Paul Kassianik, Blaine Nelson et al.NeurIPS 2024 · 835 citations
- AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language ModelsXiaogeng Liu, Nan Xu, Muhao Chen, Chaowei XiaoICLR 2024 · 722 citations
- GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via CipherYouliang Yuan, Wenxiang Jiao, Wenxuan Wang, Jen-tse Huang et al.ICLR 2024 · 441 citations
Related papers
- PAPILLON: Efficient and Stealthy Fuzz Testing-Powered Jailbreaks for LLMsXueluan Gong, Mingzhe Li, Yilin Zhang, Fengyuan Ran et al.USENIX Security 2025
- Auto-RT: Automatic Jailbreak Strategy Exploration for Red-Teaming Large Language ModelsYanjiang Liu, Shuheng Zhou, Yaojie Lu, Huijia Zhu et al.ICLR 2026 · 10 citations
- MAJIC: Markovian Adaptive Jailbreaking via Iterative Composition of Diverse Innovative StrategiesWeiwei Qi, Shuo Shao, Wei Gu, Tianhang Zheng et al.AAAI 2026
- Distract Large Language Models for Automatic Jailbreak AttackZeguan Xiao, Yan Yang, Guanhua Chen, Yun ChenEMNLP 2024 · 8 citations
- STAR: Strategy-driven Automatic Jailbreak Red-teaming For Large Language ModelJianing Liu, Qingming Li, Jiahao Chen, Rui Zeng et al.ICLR 2026 · 8 citations
