2606.11792v1 Jun 10, 2026 cs.CV

MultiToP: Learning to Patch Visual Tokens to Mitigate Hallucinations in Video Large Multimodal Models

Han Bao
Han Bao
Citations: 19
h-index: 2
Zonghui Wang
Zonghui Wang
Citations: 99
h-index: 7
Wenzhi Chen
Wenzhi Chen
Citations: 26
h-index: 2
Jiahao Yuan
Jiahao Yuan
Citations: 13
h-index: 1
Yuan Gao
Yuan Gao
Citations: 114
h-index: 4
Wenbin Xing
Wenbin Xing
Citations: 16
h-index: 2
Kaiwen Zhou
Kaiwen Zhou
Citations: 57
h-index: 2

Video Large Multimodal Models have achieved remarkable progress in video understanding, yet they remain prone to hallucinations, where generated responses are not faithfully supported by the input video. In this paper, we propose MultiToP, a multimodal-context-aware visual token patching framework that mitigates hallucinations by refining unreliable visual tokens before language generation. MultiToP introduces a lightweight Visual Token Patcher to predict token-level replacement distributions and selectively substitute unreliable visual tokens with a dynamic global patch token. To train the patcher effectively, we further propose information-guided rank calibration, which uses answer-conditioned frame-level information cues derived from the backbone to guide token replacement. Combined with ground-truth answer supervision and sparsity regularization, MultiToP enables localized visual evidence refinement without modifying the original model. Extensive experiments demonstrate that MultiToP effectively reduces hallucinations on Vript-HAL with negligible inference overhead, improving the F1 scores of Qwen3-VL-4B-Instruct by 50.60% over the vanilla model. Meanwhile, MultiToP preserves general video understanding ability, yielding an 18.58% relative accuracy gain on ActivityNet-QA for Video-LLaVA-7B.

0 Citations
0 Influential
3.5 Altmetric
17.5 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!