SEReDeEP：通过语义熵与上下文-参数融合检测检索增强模型中的幻觉
SEReDeEP: Hallucination Detection in Retrieval-Augmented Models via Semantic Entropy and Context-Parameter Fusion
AI总结：
本文针对 RAG 中外部上下文与参数知识失衡导致的幻觉问题，在 ReDeEP 基础上引入线性探针语义熵，提出 SEReDeEP 以提升幻觉评估与真值的一致性。
摘要：
检索增强生成（RAG）模型在整合外部信息与内部参数化知识时，经常遇到幻觉现象。实证研究表明，外部上下文信息与内部参数化知识之间的不平衡是幻觉产生的主要因素。现有的幻觉检测方法大多孤立地侧重外部机制或内部机制，因而忽视了二者的协同效应。最近提出的 ReDeEP 框架解耦了这两种机制，识别出导致幻觉的两个关键因素：对前馈网络（FFN）中编码的参数化知识过度依赖，以及注意力机制（尤其是复制头）对外部信息利用不足。ReDeEP 定量评估这些因素以检测幻觉，并动态调节 FFN 与复制头的贡献，以减轻幻觉发生。然而，ReDeEP 以及许多其他幻觉检测方法都被用于 logit 层面的不确定性估计或语言层面的自一致性评估，未能充分处理模型响应的语义维度，导致 RAG 实现中的幻觉评估不一致。本文在 ReDeEP 的基础上提出 SEReDeEP，通过经训练的线性探针捕获语义熵来增强计算过程，从而实现更准确反映真值评估的幻觉评估。
英文摘要：
Retrieval-Augmented Generation (RAG) models frequently encounter hallucination phenomena when integrating external information with internal parametric knowledge. Empirical studies demonstrate that the disequilibrium between external contextual information and internal parametric knowledge constitutes a primary factor in hallucination generation. Existing hallucination detection methodologies predominantly emphasize either the external or internal mechanism in isolation, thereby overlooking their synergistic effects. The recently proposed ReDeEP framework decouples these dual mechanisms, identifying two critical contributors to hallucinations: excessive reliance on parametric knowledge encoded in feed-forward networks (FFN) and insufficient utilization of external information by attention mechanisms (particularly copy heads). ReDeEP quantitatively assesses these factors to detect hallucinations and dynamically modulates the contributions of FFNs and copy heads to attenuate their occurrence. Nevertheless, ReDeEP and numerous other hallucination detection approaches have been employed at logit-level uncertainty estimation or language-level self-consistency evaluation, inadequately address the semantic dimensions of model responses, resulting in inconsistent hallucination assessments in RAG implementations. Building upon ReDeEP's foundation, this paper introduces SEReDeEP, which enhances computational processes through semantic entropy captured via trained linear probes, thereby achieving hallucination assessments that more accurately reflect ground truth evaluations.
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。
