当偏见伪装成真相：虚假相关性如何损害大语言模型中的幻觉检测
When Bias Pretends to Be Truth: How Spurious Correlations Undermine Hallucination Detection in LLMs
AI总结：
本文揭示了由训练数据中虚假相关性导致的幻觉问题，指出现有检测方法在面对此类幻觉时失效，并强调需新方法应对由此引发的幻觉。
摘要：
尽管取得了显著进展，大型语言模型（LLMs）仍然表现出幻觉现象，生成看似合理却错误的响应。在本文中，我们指出了一种关键但此前未被充分研究的幻觉类型，即由虚假相关性驱动的幻觉——训练数据中存在的特征（例如姓氏）与属性（例如国籍）之间表面但统计上显著的关联。我们证明这些虚假相关性会引发自信生成的幻觉，这些幻觉对模型规模免疫，逃避当前的检测方法，并在拒绝微调后仍然存在。通过系统控制的合成实验和对最新开源和专有LLMs（包括GPT-5）的实证评估，我们显示现有幻觉检测方法，如基于置信度的过滤和内部状态探测，在存在虚假相关性时根本无法发挥作用。我们的理论分析进一步阐明了这些统计偏见如何内在地破坏基于置信度的检测技术。我们的发现因此强调了迫切需要新的方法来专门解决由虚假相关性引起的幻觉问题。
英文摘要：
Despite substantial advances, large language models (LLMs) continue to exhibit hallucinations, generating plausible yet incorrect responses. In this paper, we highlight a critical yet previously underexplored class of hallucinations driven by spurious correlations -- superficial but statistically prominent associations between features (e.g., surnames) and attributes (e.g., nationality) present in the training data. We demonstrate that these spurious correlations induce hallucinations that are confidently generated, immune to model scaling, evade current detection methods, and persist even after refusal fine-tuning. Through systematically controlled synthetic experiments and empirical evaluations on state-of-the-art open-source and proprietary LLMs (including GPT-5), we show that existing hallucination detection methods, such as confidence-based filtering and inner-state probing, fundamentally fail in the presence of spurious correlations. Our theoretical analysis further elucidates why these statistical biases intrinsically undermine confidence-based detection techniques. Our findings thus emphasize the urgent need for new approaches explicitly designed to address hallucinations caused by spurious correlations.
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。
