芝麻开门！大型语言模型的通用黑盒越狱
Open Sesame! Universal Black Box Jailbreaking of Large Language Models
AI总结：
本文提出基于遗传算法的自动化通用黑盒越狱攻击，通过优化可与任意用户查询结合的对抗提示破坏 LLM 对齐，并经实验验证其有效性。
摘要：
大型语言模型（LLM）旨在提供有用且安全的回答，通常依赖对齐技术来适应用户意图和社会准则。遗憾的是，这种对齐可能被恶意行为者利用，他们试图操纵 LLM 的输出以达成非预期目的。本文提出一种新方法，在模型架构和参数不可访问的情况下，采用遗传算法（GA）操纵 LLM。该 GA 攻击通过优化一个通用对抗提示来工作：当该提示与用户查询结合时，会破坏被攻击模型的对齐，导致非预期且可能有害的输出。我们的新方法通过揭示模型响应偏离预期行为的实例，系统性地暴露模型的局限性和漏洞。通过大量实验，我们证明了该技术的有效性，从而为评估和增强 LLM 与人类意图的对齐提供诊断工具，为正在进行的负责任 AI 开发讨论作出贡献。据我们所知，这是首个自动化的通用黑盒越狱攻击。
英文摘要：
Large language models (LLMs), designed to provide helpful and safe responses, often rely on alignment techniques to align with user intent and social guidelines. Unfortunately, this alignment can be exploited by malicious actors seeking to manipulate an LLM's outputs for unintended purposes. In this paper we introduce a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible. The GA attack works by optimizing a universal adversarial prompt that -- when combined with a user's query -- disrupts the attacked model's alignment, resulting in unintended and potentially harmful outputs. Our novel approach systematically reveals a model's limitations and vulnerabilities by uncovering instances where its responses deviate from expected behavior. Through extensive experiments we demonstrate the efficacy of our technique, thus contributing to the ongoing discussion on responsible AI development by providing a diagnostic tool for evaluating and enhancing alignment of LLMs with human intent. To our knowledge this is the first automated universal black box jailbreak attack.
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。
