Logical Consistency as a Bridge: Improving LLM Hallucination Detection via Label Constraint Modeling between Responses and Self-Judgments
For LLM safety and reliability, this work provides a more holistic hallucination detection method by integrating neural and symbolic signals, though it is an incremental improvement over existing dual-view approaches.
LLMs hallucinate, and existing detectors use either neural uncertainty or symbolic self-judgments in isolation. LaaB bridges these by mapping symbolic labels back into the feature space via a meta-judgment process, achieving state-of-the-art hallucination detection across 4 datasets and 4 LLMs, outperforming 8 baselines.
Large Language Models (LLMs) are prone to factual hallucinations, risking their reliability in real-world applications. Existing hallucination detectors mainly extract micro-level intrinsic patterns for uncertainty quantification or elicit macro-level self-judgments through verbalized prompts. However, these methods address only a single facet of the hallucination, focusing either on implicit neural uncertainty or explicit symbolic reasoning, thereby treating these inherently coupled behaviors in isolation and failing to exploit their interdependence for a holistic view. In this paper, we propose LaaB (Logical Consistency-as-a-Bridge), a framework that bridges neural features and symbolic judgments for hallucination detection. LaaB introduces a "meta-judgment" process to map symbolic labels back into the feature space. By leveraging the inherent logical bridge where response and meta-judgment labels are either the same or opposite based on the self-judgment's semantics, LaaB aligns and integrates dual-view signals via mutual learning and enhances the hallucination detection. Extensive experiments on 4 public datasets, across 4 LLMs, against 8 baselines demonstrate the superiority of LaaB.