Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hemmat, Arshia, Torr, Philip, Chen, Yongqiang, Yu, Junchi
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2602.08048
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915784863776768
author Hemmat, Arshia
Torr, Philip
Chen, Yongqiang
Yu, Junchi
author_facet Hemmat, Arshia
Torr, Philip
Chen, Yongqiang
Yu, Junchi
contents Diffusion language models (D-LLMs) offer parallel denoising and bidirectional context, but hallucination detection for D-LLMs remains underexplored. Prior detectors developed for auto-regressive LLMs typically rely on single-pass cues and do not directly transfer to diffusion generation, where factuality evidence is distributed across the denoising trajectory and may appear, drift, or be self-corrected over time. We introduce TDGNet, a temporal dynamic graph framework that formulates hallucination detection as learning over evolving token-level attention graphs. At each denoising step, we sparsify the attention graph and update per-token memories via message passing, then apply temporal attention to aggregate trajectory-wide evidence for final prediction. Experiments on LLaDA-8B and Dream-7B across QA benchmarks show consistent AUROC improvements over output-based, latent-based, and static-graph baselines, with single-pass inference and modest overhead. These results highlight the importance of temporal reasoning on attention graphs for robust hallucination detection in diffusion language models.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08048
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TDGNet: Hallucination Detection in Diffusion Language Models via Temporal Dynamic Graphs
Hemmat, Arshia
Torr, Philip
Chen, Yongqiang
Yu, Junchi
Computation and Language
Diffusion language models (D-LLMs) offer parallel denoising and bidirectional context, but hallucination detection for D-LLMs remains underexplored. Prior detectors developed for auto-regressive LLMs typically rely on single-pass cues and do not directly transfer to diffusion generation, where factuality evidence is distributed across the denoising trajectory and may appear, drift, or be self-corrected over time. We introduce TDGNet, a temporal dynamic graph framework that formulates hallucination detection as learning over evolving token-level attention graphs. At each denoising step, we sparsify the attention graph and update per-token memories via message passing, then apply temporal attention to aggregate trajectory-wide evidence for final prediction. Experiments on LLaDA-8B and Dream-7B across QA benchmarks show consistent AUROC improvements over output-based, latent-based, and static-graph baselines, with single-pass inference and modest overhead. These results highlight the importance of temporal reasoning on attention graphs for robust hallucination detection in diffusion language models.
title TDGNet: Hallucination Detection in Diffusion Language Models via Temporal Dynamic Graphs
topic Computation and Language
url https://arxiv.org/abs/2602.08048