| title | UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal Representations |
|---|---|
| authors | Dvir Samuel, Guy Bar-Shalom, Fabrizio Frasca, Ethan Fetaya, Yftah Ziser, Gal Chechik, Haggai Maron |
| date | 2026-08-11 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.10835 |
| affiliation | NVIDIA |
| upvotes | 4 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Large Vision-Language Models (LVLMs) achieve impressive visual reasoning and dialogue capabilities, yet frequently hallucinate content unsupported by the visual input. Effective mitigation requires token-level localization, enabling targeted intervention without discarding the entire response. Existing detectors require expensive full-model fine-tuning, rely on external verifiers that ignore the model's generation process, or reduce internal signals to isolated features and hand-crafted statistics, discarding spatial, sequential, and relational structure. We introduce UniProbe, a lightweight, unified, learnable detector that models a frozen LVLM's heterogeneous computational trace from a single forward pass. UniProbe constructs a directed graph over image patches, query tokens, and generated tokens, with attention weights encoding their relations. It processes this trace with alternating structure-aware modules: a GNN for relational evidence, a ViT for 2-D visual geometry, and a GRU for response order. Interleaving them allows spatial, relational, and sequential evidence to interact throughout the detector. We further develop a streaming variant for hallucination-aware decoding, which detects and resamples hallucinated tokens during generation, and a self-adaptation strategy aligning the detector with the LVLM's own generations. Across diverse LVLM backbones, UniProbe achieves state-of-the-art token-level and object-hallucination detection. During decoding, it reduces object hallucinations by up to 55\% at 1.06times the latency of standard generation. |
cite
@misc{samuel2026uniprobe,
title={UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal Representations},
author={Dvir Samuel and Guy Bar-Shalom and Fabrizio Frasca and Ethan Fetaya and Yftah Ziser and Gal Chechik and Haggai Maron},
year={2026},
eprint={2608.10835},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.10835}
}