FRACTURED-SORRY-Bench：揭示对话轮次中攻击以削弱SORRY-Bench上拒绝效能与防御的框架（自动化多轮越狱）
FRACTURED-SORRY-Bench: Framework for Revealing Attacks in Conversational Turns Undermining Refusal Efficacy and Defenses over SORRY-Bench (Automated Multi-shot Jailbreaks)
AI总结：
本文提出FRACTURED-SORRY-Bench框架，通过将有害查询分解为无害子问题生成多轮攻击，在多个GPT模型上实现攻击成功率最高提升46.22%，揭示现有LLM安全防御的脆弱性。
摘要：
本文介绍了FRACTURED-SORRY-Bench，一个用于评估大型语言模型（LLMs）在应对多轮对话攻击时安全性的框架。基于SORRY-Bench数据集，我们提出了一种简单而有效的方法，通过将有害查询分解为看似无害的子问题来生成对抗性提示。与基线方法相比，我们的方法在GPT-4、GPT-4o、GPT-4o-mini和GPT-3.5-Turbo模型上的攻击成功率（ASRs）最大提升了+46.22%。我们证明，该技术对当前LLM安全措施构成了挑战，并强调了针对微妙多轮攻击需要更强大防御的必要性。
英文摘要：
This paper introduces FRACTURED-SORRY-Bench, a framework for evaluating the safety of Large Language Models (LLMs) against multi-turn conversational attacks. Building upon the SORRY-Bench dataset, we propose a simple yet effective method for generating adversarial prompts by breaking down harmful queries into seemingly innocuous sub-questions. Our approach achieves a maximum increase of +46.22\% in Attack Success Rates (ASRs) across GPT-4, GPT-4o, GPT-4o-mini, and GPT-3.5-Turbo models compared to baseline methods. We demonstrate that this technique poses a challenge to current LLM safety measures and highlights the need for more robust defenses against subtle, multi-turn attacks.
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。
