提出统一框架与新方法，精准识别大模型的两类幻觉。
HalluGuard: Demystifying Data-Driven and Reasoning-Driven Hallucinations in LLMs
- 基于核理论构建幻觉风险边界，分解数据与推理两类幻觉来源。
- 在10个基准上超越11种基线，9个主流模型均表现领先。
- 适合关注大模型可信性、幻觉检测的研究者与工程师。
大语言模型在医疗、法律和科学发现等高风险领域中的可靠性常因幻觉问题受损。这些错误主要源于数据驱动型幻觉与推理驱动型幻觉。现有检测方法通常仅针对单一类型，依赖任务特异性启发式规则，难以泛化至复杂场景。为此，我们提出幻觉风险边界，一种统一的理论框架，将幻觉风险形式化分解为训练时失配（数据驱动）与推理时不稳定性（推理驱动）两部分，为理解幻觉产生与演化提供原则性基础。在此基础上，我们设计了基于神经正切核（NTK）的HalluGuard评分，利用NTK诱导的几何结构与表示能力，联合识别两类幻觉。我们在10个多样化基准、11种竞争性基线及9个主流大模型上评估，结果持续达到最优性能。代码已开源：https://github.com/Susan571/HalluGuard-ICLR2026。
原文摘要 · Abstract (English)
The reliability of Large Language Models (LLMs) in high-stakes domains such as healthcare, law, and scientific discovery is often compromised by hallucinations. These failures typically stem from two sources: data-driven hallucinations and reasoning-driven hallucinations. However, existing detection methods usually address only one source and rely on task-specific heuristics, limiting their generalization to complex scenarios. To overcome these limitations, we introduce the Hallucination Risk Bound, a unified theoretical framework that formally decomposes hallucination risk into data-driven and reasoning-driven components, linked respectively to training-time mismatches and inference-time instabilities. This provides a principled foundation for analyzing how hallucinations emerge and evolve. Building on this foundation, we introduce HalluGuard, an NTK-based score that leverages the induced geometry and captured representations of the NTK to jointly identify data-driven and reasoning-driven hallucinations. We evaluate HalluGuard on 10 diverse benchmarks, 11 competitive baselines, and 9 popular LLM backbones, consistently achieving state-of-the-art performance in detecting diverse forms of LLM hallucinations. We open-source our proposed \model{} model at https://github.com/Susan571/HalluGuard-ICLR2026.
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。