Robust Multimodal Large Language Models Against Modality Conflict

2025/07/09 by Zongmeng Zhang, Zhang, Zongmeng, Wengang Zhou +5 · 2 citations
Computer Science · Engineering · #Advanced Graph Neural Networks #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Ferroelectric and Negative Capacitance Devices #Multimodal Machine Learning Applications

paper · pdf · doi:10.48550/arxiv.2507.07151

openalex publication_date 2025/07/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Despite the impressive capabilities of multimodal large language models (MLLMs) in vision-language tasks, they are prone to hallucinations in real-world scenarios. This paper investigates the hallucination phenomenon in MLLMs from the perspective of modality conflict. Unlike existing works focusing on the conflicts between model responses and inputs, we study the inherent conflicts in inputs from different modalities that place MLLMs in a dilemma and directly lead to hallucinations. We formally define the modality conflict and construct a dataset named Multimodal Modality Conflict (MMMC) to simulate this phenomenon in vision-language tasks. Three methods based on prompt engineering, supervised fine-tuning, and reinforcement learning are proposed to alleviate the hallucination caused by modality conflict. Extensive experiments are conducted on the MMMC dataset to analyze the merits and demerits of these methods. Our results show that the reinforcement learning method achieves the best performance in mitigating the hallucination under modality conflict, while the supervised fine-tuning method shows promising and stable performance. Our work sheds light on the unnoticed modality conflict that leads to hallucinations and provides more insights into the robustness of MLLMs.

Citations

• Large Vision-Language Model Alignment and Misalignment: A Survey Through the Lens of Explainability
• Insight Over Sight: Exploring the Vision-Knowledge Conflicts in Multimodal LLMs
• Unraveling Cross-Modality Knowledge Conflicts in Large Vision-Language Models
• Are We on the Right Way for Evaluating Large Vision-Language Models?
• Knowledge Conflicts for LLMs: A Survey
• RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedback
• OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation
• Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding
• Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimization
• MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI
• HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Data
• HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models
• Improved Baselines with Visual Instruction Tuning
• MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts
• Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond
• MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities
• MMBench: Is Your Multi-modal Model an All-around Player?
• Mitigating Hallucination in Large Multi-Modal Models via Robust Instruction Tuning
• MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models
• Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena
• Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts
• InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning
• Visual Instruction Tuning
• BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
• Rich Knowledge Sources Bring Complex Knowledge Conflicts: Recalibrating Models to Reflect Conflicting Evidence
• Flamingo: a Visual Language Model for Few-Shot Learning
• Training language models to follow instructions with human feedback
• Survey of Hallucination in Natural Language Generation
• Entity-Based Knowledge Conflicts in Question Answering
• Learning to summarize from human feedback
• Proximal Policy Optimization Algorithms
• A Diagram Is Worth A Dozen Images
• Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations
• REINFORCE++: Stabilizing Critic-Free Policy Optimization with Global Advantage Normalization
• OpenRLHF: An Easy-to-use, Scalable and High-performance RLHF Framework

Cited by

• FinCap: Topic-Aligned Captions for Short-Form Financial YouTube Videos
• Can Large Multimodal Models Actively Recognize Faulty Inputs? A Systematic Evaluation Framework of Their Input Scrutiny Ability

Related

• Hallucination of Multimodal Large Language Models: A Survey
• The Curse of Multi-Modalities: Evaluating Hallucinations of Large Multimodal Models across Language, Visual, and Audio
• Hallucination Augmented Contrastive Learning for Multimodal Large Language Model
• Modality Bias in LVLMs: Analyzing and Mitigating Object Hallucination via Attention Lens
• Investigating and Mitigating the Multimodal Hallucination Snowballing in Large Vision-Language Models