Year: 2026
Type: conference-paper
Abstract: Large Language Models (LLMs) frequently "hallucinate" plausible but incorrect assertions, a vulnerability often missed by uncertainty metrics when models are "confidently wrong."We propose DiffuTruth, an unsupervised framework that re-conceptualizes fact verification via non-equilibrium thermodynamics, positing that factual truths act as stable attractors on a generative manifold while hallucinations are unstable.We introduce the "Generative Stress Test": claims are corrupted with noise and reconstructed using a discrete text diffusion model.We define Semantic Energy, a metric measuring the semantic divergence between the original claim and its reconstruction using an NLI critic.Unlike vector-space errors, Semantic Energy isolates deep factual contradictions.We further propose a Hybrid Calibration fusing this stability signal with discriminative confidence.Extensive experiments on FEVER demonstrate DiffuTruth achieves a state-of-theart unsupervised AUROC of 0.725, outperforming baselines by +1.5% through the correction of overconfident predictions.Furthermore, we show superior zero-shot generalization on the multi-hop HOVER dataset, outperforming baselines by over 4%, confirming the robustness of thermodynamic truth properties to distribution shifts.
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7140145192