用最大均值差异检测大模型幻觉，无需真实标签。
MMD-Flagger: Leveraging Maximum Mean Discrepancy to Detect Hallucinations
- 通过分析不同温度下输出的分布变化，追踪MMD轨迹
- 在多语言验证集上对齐主流模型，识别幻觉行为
- 适合构建无标注场景下的可信智能体系统
大型语言模型（LLMs）正日益融入代理型AI系统，但其生成幻觉的问题仍是关键安全挑战。在测试时检测这些事实性错误，尤其在缺乏真实标签的情况下，对于构建可信自主代理至关重要。我们提出MMD-Flagger，一种利用最大均值差异（MMD）的方法，通过监控不同解码温度下LLM输出的稳定性来检测幻觉。该方法跟踪某一解码配置下的模型输出与一组随机样本之间的MMD轨迹，依据轨迹的特征形状识别幻觉。我们在多语言声明验证基准（MUCH）上评估了MMD-Flagger，使用Llama-3系列和Gemma-3等现代大模型。
原文摘要 · Abstract (English)
Large Language Models (LLMs) are increasingly integrated into agentic AI systems, yet their propensity to generate hallucinations remains a critical safety concern. Detecting these factual errors at test-time, particularly without ground-truth labels, is essential for building trustworthy autonomous agents. We propose MMD-Flagger, an hallucination detection method that utilizes Maximum Mean Discrepancy (MMD) and monitors the stability of LLM outputs across varying decoding temperatures. Our method tracks the MMD trajectory between a LLM's response at a certain decoding configuration and a set of stochastic samples, identifying hallucinations based on the trajectory's characteristic shape. We evaluate MMDFlagger on multi-lingual claim verification benchmarks (MUCH) using modern LLMs like Llama-3 families and Gemma-3.
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。