MUSE通过强化搜索机制，同时攻防多轮对话中的安全漏洞。
MUSE: MCTS-Driven Red Teaming Framework for Enhanced Multi-Turn Dialogue Safety in Large Language Models
- 用语义框架与启发式搜索探索多样对话路径
- 在多个模型上验证可有效发现并缓解多轮攻击
- 适合研究大模型安全或对抗攻防的开发者
随着大语言模型（LLMs）广泛应用，确保其与人类价值观对齐至关重要，以防止攻击者通过操纵模型生成有害内容。现有防御多针对单轮攻击，而真实场景常涉及多轮对话，攻击者可利用对话上下文绕过安全机制。本文提出MUSE框架，从攻防双重视角应对多轮越狱攻击。攻击端引入MUSE-A，基于语义框架与启发式树搜索，探索多样化的语义轨迹；防御端提出MUSE-D，采用细粒度安全对齐策略，在对话早期介入，降低模型脆弱性。在多种模型上的大量实验表明，MUSE能有效识别并缓解多轮安全漏洞。代码已公开于https://github.com/yansiyu02/MUSE。
原文摘要 · Abstract (English)
As large language models~(LLMs) become widely adopted, ensuring their alignment with human values is crucial to prevent jailbreaks where adversaries manipulate models to produce harmful content. While most defenses target single-turn attacks, real-world usage often involves multi-turn dialogues, exposing models to attacks that exploit conversational context to bypass safety measures. We introduce MUSE, a comprehensive framework tackling multi-turn jailbreaks from both attack and defense angles. For attacks, we propose MUSE-A, a method that uses frame semantics and heuristic tree search to explore diverse semantic trajectories. For defense, we present MUSE-D, a fine-grained safety alignment approach that intervenes early in dialogues to reduce vulnerabilities. Extensive experiments on various models show that MUSE effectively identifies and mitigates multi-turn vulnerabilities. Code is available at \href{https://github.com/yansiyu02/MUSE}{https://github.com/yansiyu02/MUSE}.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。