登门槛效应:从认知心理学视角理解大语言模型越狱
Foot In The Door: Understanding Large Language Model Jailbreaking via Cognitive Psychology
AI总结:
本文从认知心理学视角解释大语言模型越狱机制,提出基于登门槛效应的自动黑盒越狱方法,通过多步增量提示逐步诱导模型回答有害问题,在8个先进LLMs上取得83.9%的平均成功率。
摘要:
大语言模型(LLMs)已逐渐成为人们获取新知识的入口。然而,攻击者可以打破模型的安全防护(“监狱”)以访问受限信息,这被称为“越狱”。此前的研究已经表明,当前LLMs在面对此类越狱攻击时存在弱点。尽管如此,对于LLMs在收到越狱提示时内在决策机制的理解明显不足。我们的研究为越狱提示提供了心理学解释。借鉴认知一致性理论,我们认为越狱的关键在于引导LLM在错误方向上实现认知协调。进一步,我们提出了一种基于登门槛(FITD)技术的自动黑盒越狱方法。该方法通过多步增量提示逐步诱导模型回答有害问题。我们实例化了一个原型系统,在8个先进LLMs上评估越狱有效性,取得了83.9%的平均成功率。本研究从心理学视角构建了对LLMs内在决策逻辑的解释性洞见。
英文摘要:
Large Language Models (LLMs) have gradually become the gateway for people to acquire new knowledge. However, attackers can break the model's security protection ("jail") to access restricted information, which is called "jailbreaking." Previous studies have shown the weakness of current LLMs when confronted with such jailbreaking attacks. Nevertheless, comprehension of the intrinsic decision-making mechanism within the LLMs upon receipt of jailbreak prompts is noticeably lacking. Our research provides a psychological explanation of the jailbreak prompts. Drawing on cognitive consistency theory, we argue that the key to jailbreak is guiding the LLM to achieve cognitive coordination in an erroneous direction. Further, we propose an automatic black-box jailbreaking method based on the Foot-in-the-Door (FITD) technique. This method progressively induces the model to answer harmful questions via multi-step incremental prompts. We instantiated a prototype system to evaluate the jailbreaking effectiveness on 8 advanced LLMs, yielding an average success rate of 83.9%. This study builds a psychological perspective on the explanatory insights into the intrinsic decision-making logic of LLMs.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。