# Mitigating Hallucinations in Large Vision-Language Models via DPO: On-Policy Data Hold the Key

arXiv: https://arxiv.org/abs/2501.09695 · HTML: https://mlsanity.com/paper/2501.09695/

## Key facts

- Cited by 65 papers in the map (data through 14 Oct 2026).
- Placed in Hallucination mitigation in LVLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 48 of its 64 references that are in the map.
- First posted on arXiv 16 Jan 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Hallucination mitigation in LVLMs](https://mlsanity.com/mllm/hallucination-mitigation-in-lvlms/)
- LLM alignment / RLHF / preference optimisation / [Diffusion and multimodal model alignment](https://mlsanity.com/alignment/diffusion-and-multimodal-model-alignment/)

## Builds on (most cited first, 30 of 48)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [Fine-Tuning Language Models from Human Preferences](https://mlsanity.com/paper/1909.08593/) (2019)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems](https://mlsanity.com/paper/2005.01643/) (2020)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day](https://mlsanity.com/paper/2306.00890/) (2023)
- [SimPO: Simple Preference Optimization with a Reference-Free Reward](https://mlsanity.com/paper/2405.14734/) (2024)
- [Aligning Large Multimodal Models with Factually Augmented RLHF](https://mlsanity.com/paper/2309.14525/) (2023)
- [Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding](https://mlsanity.com/paper/2311.16922/) (2023)
- [Object Hallucination in Image Captioning](https://mlsanity.com/paper/1809.02156/) (2018)
- [Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models](https://mlsanity.com/paper/2311.07919/) (2023)
- [ORPO: Monolithic Preference Optimization without Reference Model](https://mlsanity.com/paper/2403.07691/) (2024)
- [OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation](https://mlsanity.com/paper/2311.17911/) (2023)
- [RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedback](https://mlsanity.com/paper/2312.00849/) (2023)
- [Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation](https://mlsanity.com/paper/2401.08417/) (2024)

## Built on by (most cited first, 30 of 65)

- [HunyuanVideo 1.5 Technical Report](https://mlsanity.com/paper/2511.18870/) (2025)
- [Fine-Grained Preference Optimization Improves Spatial Reasoning in VLMs](https://mlsanity.com/paper/2506.21656/) (2025)
- [Learning Dynamics of VLM Finetuning](https://mlsanity.com/paper/2510.11978/) (2025)
- [Mitigating Hallucination Through Theory-Consistent Symmetric Multimodal Preference Optimization](https://mlsanity.com/paper/2506.11712/) (2025)
- [Aligning Multimodal LLM with Human Preference: A Survey](https://mlsanity.com/paper/2503.14504/) (2025)
- [Mitigating Hallucination in VideoLLMs via Temporal-Aware Activation Engineering](https://mlsanity.com/paper/2505.12826/) (2025)
- [Not All Tokens and Heads Are Equally Important: Dual-Level Attention Intervention for Hallucination Mitigation](https://mlsanity.com/paper/2506.12609/) (2025)
- [Look Carefully: Adaptive Visual Reinforcements in Multimodal Large Language Models for Hallucination Mitigation](https://mlsanity.com/paper/2602.24041/) (2026)
- [MoD-DPO: Towards Mitigating Cross-modal Hallucinations in Omni LLMs using Modality Decoupled Preference Optimization](https://mlsanity.com/paper/2603.03192/) (2026)
- [UniVRSE: Unified Vision-conditioned Response Semantic Entropy for Hallucination Detection in Medical Vision-Language Models](https://mlsanity.com/paper/2503.20504/) (2025)
- [CheXPO: Preference Optimization for Chest X-ray VLMs with Counterfactual Rationale](https://mlsanity.com/paper/2507.06959/) (2025)
- [Toward Clinically Explainable AI for Medical Diagnosis: A Foundation Model with Human-Compatible Reasoning via Reinforcement Learning](https://mlsanity.com/paper/2509.03906/) (2025)
- [Residual Decoding: Mitigating Hallucinations in Large Vision-Language Models via History-Aware Residual Guidance](https://mlsanity.com/paper/2602.01047/) (2026)
- [Seeing Through the Chain: Mitigate Hallucination in Multimodal Reasoning Models via CoT Compression and Contrastive Preference Optimization](https://mlsanity.com/paper/2602.03380/) (2026)
- [KVSmooth: Mitigating Hallucination in Multi-modal Large Language Models through Key-Value Smoothing](https://mlsanity.com/paper/2602.04268/) (2026)
- [Fighting Hallucinations with Counterfactuals: Diffusion-Guided Perturbations for LVLM Hallucination Suppression](https://mlsanity.com/paper/2603.10470/) (2026)
- [Visual Preference Optimization with Rubric Rewards](https://mlsanity.com/paper/2604.13029/) (2026)
- [Reallocating Attention Across Layers to Reduce Multimodal Hallucination](https://mlsanity.com/paper/2510.10285/) (2025)
- [Optimizing LVLMs with On-Policy Data for Effective Hallucination Mitigation](https://mlsanity.com/paper/2512.00706/) (2025)
- [Revis: Sparse Latent Steering to Mitigate Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2602.11824/) (2026)
- [FINER: MLLMs Hallucinate under Fine-grained Negative Queries](https://mlsanity.com/paper/2603.17662/) (2026)
- [Same Attention, Different Truths: Put Logit-Lens over Visual Attention to Detect and Mitigate LVLM Object Hallucination](https://mlsanity.com/paper/2608.07302/) (2026)
- [RL makes MLLMs see better than SFT](https://mlsanity.com/paper/2510.16333/) (2025)
- [Reflective Preference Optimization (RPO): Enhancing On-Policy Alignment via Hint-Guided Reflection](https://mlsanity.com/paper/2512.13240/) (2025)
- [DA-DPO: Cost-efficient Difficulty-aware Preference Optimization for Reducing MLLM Hallucinations](https://mlsanity.com/paper/2601.00623/) (2026)
- [Do Understanding and Generation Fight? A Diagnostic Study of DPO for Unified Multimodal Models](https://mlsanity.com/paper/2603.17044/) (2026)
- [SHAPE : Self-Improved Visual Preference Alignment by Iteratively Generating Holistic Winner](https://mlsanity.com/paper/2503.04858/) (2025)
- [OViP: Online Vision-Language Preference Learning for VLM Hallucination](https://mlsanity.com/paper/2505.15963/) (2025)
- [Aligning VLM Assistants with Personalized Situated Cognition](https://mlsanity.com/paper/2506.00930/) (2025)
- [Anatomical Region-Guided Contrastive Decoding: A Plug-and-Play Strategy for Mitigating Hallucinations in Medical VLMs](https://mlsanity.com/paper/2512.17189/) (2025)

## Abstract (arXiv)

Hallucination remains a major challenge for Large Vision-Language Models (LVLMs). Direct Preference Optimization (DPO) has gained increasing attention as a simple solution to hallucination issues. It directly learns from constructed preference pairs that reflect the severity of hallucinations in responses to the same prompt and image. Nonetheless, different data construction methods in existing works bring notable performance variations. We identify a crucial factor here: outcomes are largely contingent on whether the constructed data aligns on-policy w.r.t the initial (reference) policy of DPO. Theoretical analysis suggests that learning from off-policy data is impeded by the presence of KL-divergence between the updated policy and the reference policy. From the perspective of dataset distribution, we systematically summarize the inherent flaws in existing algorithms that employ DPO to address hallucination issues. To alleviate the problems, we propose On-Policy Alignment (OPA)-DPO framework, which uniquely leverages expert feedback to correct hallucinated responses and aligns both the original and expert-revised responses in an on-policy manner. Notably, with only 4.8k data, OPA-DPO achieves an additional reduction in the hallucination rate of LLaVA-1.5-7B: 13.26% on the AMBER benchmark and 5.39% on the Object-Hal benchmark, compared to the previous SOTA algorithm trained with 16k samples. Our implementation is available at https://github.com/zhyang2226/OPA-DPO.