cs.CVNov 3, 2026

Entropy-Gated Dynamic Rectification for Object Hallucination in Large Vision-Language Models

Authors: Riasad Alvi, Md. Asaduzzaman Shuvo, Md. Ahashan Habib, Md. Faiyaz Abdullah Sayeedi

Organizations: United International University, Bangladesh · BRAC University, Bangladesh

Abstract

Large Vision-Language Models (LVLMs) routinely name objects that are absent from the image. Latent steering can reduce this failure in one model forward pass, but a fixed gate treats early and late decoding steps by the same criterion. We observe that predictive entropy varies sharply within free-running generation, while states that commit to short answers occupy a narrow, confident band. We exploit this temporal mismatch by using uncertainty as a causally available, token-local control signal for an existing steering gate. We propose Entropy-Gated Dynamic Rectification (EGDR), which lowers the intervention threshold when entropy derived from the preceding step's already available logits is high and retains the calibrated operating point when it is low. On LLaVA-1.5-7B, the underlying static gate lowers CHAIR CS/CI from 53.00/15.30 to 32.00/14.30; replacing only its constant threshold with EGDR further lowers them to 29.00/13.60. At a matched 58% intervention rate, the corresponding static control obtains 30.50/14.00. The gain over the static gate holds on LLaVA-NeXT-7B and grows with the generation budget, while POPE scores and general multimodal capability remain effectively unchanged. EGDR requires no additional model forward pass and adds roughly 2% measured latency. Our code is available at https://github.com/alvi-uiu/EGDR.

Figures & tables

Explore similar work

CardsList
  1. Mitigating Entangled Steering in Large Vision-Language Models for Hallucination Reduction

    Apr 9, 2026Yuanhong Zhang, Zhaoyang Wang, Xin Zhang +2activation steering in vision-language modelsVLM Hallucination

  2. NoLan: Mitigating Object Hallucinations in Large Vision-Language Models via Dynamic Suppression of Language Priors

    Feb 25, 2026Lingfeng Ren, Weihao Yu, Runpeng Yu +1Object Hallucination in VLMsVLM Hallucination Mitigation