HalluSAE: Detecting Hallucinations in Large Language Models via Sparse Auto-Encoders
For LLM safety, this provides a principled detection method grounded in model dynamics, but is incremental over existing sparse autoencoder approaches.
HalluSAE models hallucination in LLMs as a phase transition in latent dynamics, using sparse autoencoders to detect critical shifts. It achieves state-of-the-art detection on Gemma-2-9B, outperforming prior methods.
Large Language Models (LLMs) are powerful and widely adopted, but their practical impact is limited by the well-known hallucination phenomenon. While recent hallucination detection methods have made notable progress, we find most of them overlook the dynamic nature and underlying mechanisms of it. To address this gap, we propose HalluSAE, a phase transition-inspired framework that models hallucination as a critical shift in the model's latent dynamics. By modeling the generation process as a trajectory through a potential energy landscape, HalluSAE identifies critical transition zones and attributes factual errors to specific high-energy sparse features. Our approach consists of three stages: (1) Potential Energy Empowered Phase Zone Localization via sparse autoencoders and a geometric potential energy metric; (2) Hallucination-related Sparse Feature Attribution using contrastive logit attribution; and (3) Probing-based Causal Hallucination Detection through linear probes on disentangled features. Extensive experiments on Gemma-2-9B demonstrate that HalluSAE achieves state-of-the-art hallucination detection performance.