mlsanityData to 7 Oct 2026

Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMs

arXiv 2606.28401 ↗First posted 24 Jun 2026cs.CV cs.LGcited by 0

On the map: Multimodal LLMs / vision-language models → Hallucination mitigation in LVLMs

also in 1 more
Abstract (arXiv)

Vision-Language Models (VLMs) have shown strong performance in visual understanding, yet they still suffer from hallucinations, generating content that is not grounded in the image. Preference alignment is a promising approach to improve visual faithfulness, but its success depends heavily on how preference pairs are constructed. Existing methods exhibit two key limitations; (a) intervention-based methods often introduce significant deviation from the policy distribution, and (b) sampling-based methods often underuse visual information during the construction. In this paper, we propose ViPSy (Vision-driven Preference Synthesis), a framework for constructing preference data that are both policy-aligned and visually grounded. Our framework consists of two stages; in the first stage, ViPSy derives a visual cue from recurring object-level content across semantically aligned image variants, so preference construction can rely on visual information rather than language priors. In the second stage, ViPSy conditions the policy's own rollouts on this cue, allowing candidates to be guided by visually grounded content while staying close to the policy's response distribution. The resulting candidates remain close to the policy's response distribution while better leveraging visual information from the image. Experiments show that the resulting VLM, preference-aligned with ViPSy-constructed preference pairs, achieves a new state-of-the-art in hallucination mitigation. Compared with the previous state-of-the-art method, it reduces hallucination rates on AMBER and Object HalBench by 35.7% and 24.5%, respectively. The resulting model further improves on general visual grounding benchmarks, e.g., MMStar, MMVP, and CV-Bench, while also yielding gains in semantic segmentation and ImageNet linear probing, underscoring the effectiveness of our framework in enhancing the model's visual capabilities.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Hallucination mitigation in LVLMs, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 86 of its 96 references that are in the map. First posted on arXiv 24 Jun 2026.

Builds on

86 of 96 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025SHIELD: Suppressing Hallucinations In LVLM Encoders via Bias and Vulnerability Defense18 Oct
  2. 2025Mitigating Object Hallucinations via Sentence-Level Early Intervention16 Jul
  3. 2025Mitigating Hallucination Through Theory-Consistent Symmetric Multimodal Preference Optimization13 Jun
  4. 2025Mitigating Hallucinations in Large Vision-Language Models via DPO: On-Policy Data Hold the Key16 Jan
  5. 2024Systematic Reward Gap Optimization for Mitigating VLM Hallucinations26 Nov
  6. 2024V-DPO: Mitigating Hallucination in Large Vision Language Models via Vision-Guided Direct Preference Optimization5 Nov
  7. 2024Mitigating Object Hallucination in MLLMs via Data-augmented Phrase-level Alignment28 May
  8. 2024Detecting and Mitigating Hallucination in Large Vision Language Models via Fine-Grained AI Feedback22 Apr
  9. 2024EFUF: Efficient Fine-grained Unlearning Framework for Mitigating Hallucinations in Multimodal Large Language Models15 Feb
  10. 2023RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedback1 Dec
  11. 2023OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation29 Nov
  12. 2023Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding28 Nov
4 more
  1. 2023AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluation13 Nov
  2. 2023Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision13 Nov
  3. 2023HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models23 Oct
  4. 2023Detecting and Preventing Hallucinations in Large Vision Language Models11 Aug

Other references · 70

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  3. 2022Training language models to follow instructions with human feedbackcited by 15,712
  4. 2017Proximal Policy Optimization Algorithmscited by 14,441
  5. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  6. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  7. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  8. 2023Improved Baselines with Visual Instruction Tuningcited by 4,362
  9. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,337
  10. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  11. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
  12. 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,274
38 more
  1. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,048
  2. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  3. 2023InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuningcited by 2,855
  4. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  5. 2016Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotationscited by 2,646
  6. 2020Learning to summarize from human feedbackcited by 2,562
  7. 2025Qwen3-VL Technical Reportcited by 2,534
  8. 2016Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answeringcited by 2,498
  9. 2019Fine-Tuning Language Models from Human Preferencescited by 2,124
  10. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contextscited by 1,582
  11. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,419
  12. 2019Natural Adversarial Examplescited by 1,401
  13. 2023LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Daycited by 1,287
  14. 2024Qwen2.5-Coder Technical Reportcited by 1,151
  15. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 951
  16. 2025OpenAI GPT-5 System Cardcited by 910
  17. 2023A General Theoretical Paradigm to Understand Learning from Human Preferencescited by 906
  18. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 893
  19. 2023On-Policy Distillation of Language Models: Learning from Self-Generated Mistakescited by 840
  20. 2024Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMscited by 822
  21. 2023Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedbackcited by 743
  22. 2024Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMscited by 719
  23. 2023Aligning Large Multimodal Models with Factually Augmented RLHFcited by 703
  24. 2024Self-Rewarding Language Modelscited by 654
  25. 2024ORPO: Monolithic Preference Optimization without Reference Modelcited by 556
  26. 2025Visual-RFT: Visual Reinforcement Fine-Tuningcited by 501
  27. 2024YOLO-World: Real-Time Open-Vocabulary Object Detectioncited by 498
  28. 2023SLiC-HF: Sequence Likelihood Calibration with Human Feedbackcited by 395
  29. 2023Med-Flamingo: a Multimodal Medical Few-shot Learnercited by 387
  30. 2026Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Modelscited by 386
  31. 2023DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Modelscited by 367
  32. 2023WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentationcited by 311
  33. 2024mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Modelscited by 285
  34. 2024Self-Play Preference Optimization for Language Model Alignmentcited by 228
  35. 2024Aligning Modalities in Vision Large Language Models via Preference Fine-tuningcited by 197
  36. 2024Preference Fine-Tuning of LLMs Should Leverage Suboptimal, On-Policy Datacited by 184
  37. 2023AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Modelscited by 169
  38. 2022Omni3D: A Large Benchmark and Model for 3D Object Detection in the Wildcited by 159

50 most cited of 70 All 70 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Mitigating Hallucinations in Large Vision-Language Models via Entity-Centric Multimodal Preference Optimization
  2. 2025OViP: Online Vision-Language Preference Learning for VLM Hallucination
  3. 2025Mitigating Visual Hallucinations via Semantic Curriculum Preference Optimization in MLLMs
  4. 2026Aligning with Your Own Voice: Self-Corrected Preference Learning for Hallucination Mitigation in LVLMs
  5. 2025CHiP: Cross-modal Hierarchical Direct Preference Optimization for Multimodal LLMs
  6. 2023Silkie: Preference Distillation for Large Visual Language Models
  7. 2024CLIP-DPO: Vision-Language Models as a Source of Preference for Fixing Hallucinations in LVLMs
  8. 2024Cracking the Code of Hallucination in LVLMs with Vision-aware Head Divergence