arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2606.01441cs.AI
收藏

深入歧义:基于A*的多智能体常识混淆攻击LLM提示

Dive into Ambiguity: A*-Inspired Multi-Agents Commonsense Obfuscation Attack on LLM Prompts

University of Liverpool(利物浦大学)

Boxuan Wang, Zhuoyun Li, Xiaowei Huang, Yi Dong

AI总结:

提出一种基于A*的事实错误诱导框架,通过层次化重写策略和动态语义分散系数生成语义对齐但混淆的提示,以高效攻击LLM的常识推理。

摘要:

大型语言模型(LLM)在推理和知识密集型任务中表现出色,但仍易受到保留意图同时触发常识幻觉的提示级对抗攻击。这一漏洞亟待解决,因为LLM正迅速集成到事实可靠性不容妥协的安全关键领域。现有攻击方法要么缺乏效率,要么无法捕捉真实世界对手的适应性策略。我们提出一种基于A*的事实错误诱导框架,用于生成语义对齐但混淆的提示。其核心是由动态语义分散系数$γ$引导的层次化重写策略,该系数遵循反向模拟退火调度,在早期平衡保守编辑,后期进行激进混淆。为了增强可解释性,我们进一步引入智能体机制标记,发现并优化对抗机制,提供可解释的反向优化。理论上,我们证明提示重写遵循收缩递归,导致随着$γ$减小语义崩溃。实验上,在多种LLM上,我们的方法以更少的尝试次数实现了比穷举探索更高的攻击成功率,证明了其高效性和有效性。

英文摘要:

Large language models (LLMs) excel in reasoning and knowledge-intensive tasks but remain vulnerable to prompt-level adversarial attacks that preserve intent while triggering commonsense hallucinations. This vulnerability is urgent, as LLMs are rapidly integrated into safety-critical domains where factual reliability is non-negotiable. Existing attack methods either lack efficiency or fail to capture the adaptive strategies of real-world adversaries. We propose an A*-inspired Factual Error Induction Framework, a framework for generating semantically aligned yet obfuscated prompts. At its core is a Hierarchical Rewrite Strategy guided by a dynamic semantic dispersion coefficient $γ$ that balances conservative edits early with aggressive obfuscations later, following a reverse simulated annealing schedule. To enhance interpretability, we further introduce Agentic Mechanism Labeling, which discovers and refines adversarial mechanisms, offering interpretable reverse optimization. Theoretically, we prove that prompt rewriting follows a contractive recurrence, leading to semantic collapse as $γ$ decreases. Empirically, across diverse LLMs, our method achieves higher attack success rates than exhaustive exploration while requiring fewer attempts, demonstrating both efficiency and effectiveness.

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。

↑