Look Carefully: Adaptive Visual Reinforcements in Multimodal Large Language Models for Hallucination Mitigation

TL;DR

This paper introduces AIR, a training-free, adaptive visual reinforcement method for multimodal large language models that reduces hallucinations by selectively emphasizing salient visual cues during decoding.

Contribution

AIR is a novel, training-free framework that improves hallucination mitigation in MLLMs through prototype-based token reduction and OT-guided patch reinforcement.

Findings

Significantly reduces hallucinations in MLLMs

Maintains core reasoning and understanding capabilities

Effective across multiple MLLMs and tasks

Abstract

Multimodal large language models (MLLMs) have achieved remarkable progress in vision-language reasoning, yet they remain vulnerable to hallucination, where generated content deviates from visual evidence. Existing mitigation strategies either require costly supervision during training or introduce additional latency at inference time. Recent vision enhancement methods attempt to address this issue by reinforcing visual tokens during decoding, but they typically inject all tokens indiscriminately, which causes interference from background regions and distracts the model from critical cues. To overcome this challenge, we propose Adaptive Visual Reinforcement (AIR), a training-free framework for MLLMs. AIR consists of two components. Prototype-based token reduction condenses the large pool of visual tokens into a compact subset to suppress redundancy. OT-guided patch reinforcement quantifies the alignment between hidden states and patch embeddings to selectively integrate the most consistent patches into feed-forward layers. As a result, AIR enhances the model's reliance on salient visual information and effectively mitigates hallucination. Extensive experiments across representative MLLMs demonstrate that AIR substantially reduces hallucination while preserving general capabilities, establishing it as an effective solution for building reliable MLLMs.