PAPILLON: Efficient and Stealthy Fuzz Testing-Powered Jailbreaks for LLMs
This addresses the vulnerability of LLMs to scalable and stealthy jailbreak attacks, which is a critical security issue for AI developers and users, representing a novel method rather than an incremental improvement.
The paper tackles the problem of jailbreaking LLMs by introducing PAPILLON, an automated black-box framework that uses fuzz testing and customized designs to generate stealthy and efficient jailbreak prompts, achieving attack success rates over 90% on GPT-3.5 turbo and reducing prompt length to 100 tokens while maintaining high coherence.
Large Language Models (LLMs) have excelled in various tasks but are still vulnerable to jailbreaking attacks, where attackers create jailbreak prompts to mislead the model to produce harmful or offensive content. Current jailbreak methods either rely heavily on manually crafted templates, which pose challenges in scalability and adaptability, or struggle to generate semantically coherent prompts, making them easy to detect. Additionally, most existing approaches involve lengthy prompts, leading to higher query costs. In this paper, to remedy these challenges, we introduce a novel jailbreaking attack framework called PAPILLON, which is an automated, black-box jailbreaking attack framework that adapts the black-box fuzz testing approach with a series of customized designs. Instead of relying on manually crafted templates,PAPILLON starts with an empty seed pool, removing the need to search for any related jailbreaking templates. We also develop three novel question-dependent mutation strategies using an LLM helper to generate prompts that maintain semantic coherence while significantly reducing their length. Additionally, we implement a two-level judge module to accurately detect genuine successful jailbreaks. We evaluated PAPILLON on 7 representative LLMs and compared it with 5 state-of-the-art jailbreaking attack strategies. For proprietary LLM APIs, such as GPT-3.5 turbo, GPT-4, and Gemini-Pro, PAPILLONs achieves attack success rates of over 90%, 80%, and 74%, respectively, exceeding existing baselines by more than 60\%. Additionally, PAPILLON can maintain high semantic coherence while significantly reducing the length of jailbreak prompts. When targeting GPT-4, PAPILLON can achieve over 78% attack success rate even with 100 tokens. Moreover, PAPILLON demonstrates transferability and is robust to state-of-the-art defenses. Code: https://github.com/aaFrostnova/Papillon