PoLLMgraph：通过状态转移动力学揭示大型语言模型中的幻觉
PoLLMgraph: Unraveling Hallucinations in Large Language Models via State Transition Dynamics
AI总结：
本文提出白盒方法 PoLLMgraph，利用可处理概率模型分析 LLM 生成时的内部状态转移动力学以检测和预测幻觉，并在 TruthfulQA 等基准上实现超过 20% 的 AUC-ROC 提升。
摘要：
尽管近年来大型语言模型（LLM）取得了巨大进步，但其实际部署面临的一个尤为紧迫的挑战是幻觉现象，即模型捏造事实并生成非事实性陈述。为此，我们提出 PoLLMgraph，一种面向 LLM 的测谎仪，作为有效的基于模型的白盒检测与预测方法。PoLLMgraph 与大量现有研究明显不同，后者主要通过黑盒评估来应对此类挑战。具体而言，我们证明，通过使用可处理概率模型分析 LLM 在生成过程中的内部状态转移动力学，可以有效检测幻觉。在多种开源 LLM 上的实验结果证实了 PoLLMgraph 的有效性，其性能大幅优于最先进方法；在 TruthfulQA 等常用基准数据集上，AUC-ROC 提升超过 20%。我们的工作为基于模型的 LLM 白盒分析开辟了新途径，激励研究社区进一步探索、理解并完善 LLM 行为中复杂的动力学。
英文摘要：
Despite tremendous advancements in large language models (LLMs) over recent years, a notably urgent challenge for their practical deployment is the phenomenon of hallucination, where the model fabricates facts and produces non-factual statements. In response, we propose PoLLMgraph, a Polygraph for LLMs, as an effective model-based white-box detection and forecasting approach. PoLLMgraph distinctly differs from the large body of existing research that concentrates on addressing such challenges through black-box evaluations. In particular, we demonstrate that hallucination can be effectively detected by analyzing the LLM's internal state transition dynamics during generation via tractable probabilistic models. Experimental results on various open-source LLMs confirm the efficacy of PoLLMgraph, outperforming state-of-the-art methods by a considerable margin, evidenced by over 20% improvement in AUC-ROC on common benchmarking datasets like TruthfulQA. Our work paves a new way for model-based white-box analysis of LLMs, motivating the research community to further explore, understand, and refine the intricate dynamics of LLM behaviors.
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。
