mlsanityData to 7 Oct 2026

SAVOR: Self-Aware Visual Grounding via Confidence-Calibrated Reinforcement Learning for Multimodal Hallucination Mitigation

arXiv 2609.16601 ↗First posted 15 Sep 2026cs.CVcited by 0

On the map: Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

Abstract (arXiv)

Multimodal large language models (MLLMs) have made strong progress on visual question answering and image captioning, yet they still produce fluent claims about objects, attributes, or relations that are not grounded in the image. Many remedies either modify decoding at test time, which adds latency, or fine tune with preferences such as DPO variants, which teach which answer is preferred but not when the model's own answer is unreliable. We argue that calibrated self assessment is the missing signal. We introduce Savor, a training framework that (i) augments the output schema with token and answer confidence, (ii) optimises the policy with a Group Relative Policy Optimisation (GRPO) objective that penalises calibration error and poor abstention decisions, and (iii) uses the learned confidence at inference time to revisit visual evidence only when the model is uncertain. Experiments on POPE, HallusionBench, AMBER and MMHal-Bench across two recent backbones (InternVL3-8B and Qwen3-VL-8B) show that Savor reduces hallucination while preserving general capability on MME and MMBench, with lower Expected Calibration Error than DPO and decoding baselines.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 53 of its 56 references that are in the map. First posted on arXiv 15 Sep 2026.

Builds on

53 of 56 references.

References · 53

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  4. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  5. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  6. 2017On Calibration of Modern Neural Networkscited by 4,860
  7. 2025Qwen3-VL Technical Reportcited by 2,534
  8. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  9. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,769
  10. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,696
  11. 2023Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedbackcited by 788
  12. 2022Teaching Models to Express Their Uncertainty in Wordscited by 731
38 more
  1. 2023Aligning Large Multimodal Models with Factually Augmented RLHFcited by 703
  2. 2023Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decodingcited by 671
  3. 2018Object Hallucination in Image Captioningcited by 661
  4. 2024Self-Rewarding Language Modelscited by 654
  5. 2023HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Modelscited by 572
  6. 2023OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocationcited by 468
  7. 2023RLHF-V: Towards Trustworthy MLLMs via Behavior Alignment from Fine-grained Correctional Human Feedbackcited by 404
  8. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learningcited by 315
  9. 2023AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluationcited by 293
  10. 2023Detecting and Preventing Hallucinations in Large Vision Language Modelscited by 254
  11. 2025LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Trainingcited by 233
  12. 2023Woodpecker: Hallucination Correction for Multimodal Large Language Modelscited by 223
  13. 2023Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimizationcited by 223
  14. 2024HALC: Object Hallucination Reduction via Adaptive Focal-Contrast Decodingcited by 171
  15. 2024SaySelf: Teaching LLMs to Express Confidence with Self-Reflective Rationalescited by 102
  16. 2024mDPO: Conditional Preference Optimization for Multimodal Large Language Modelscited by 92
  17. 2024Reducing Hallucinations in Vision-Language Models via Latent Space Steeringcited by 56
  18. 2025Seeing Across Views: Benchmarking Spatial Reasoning of Vision-Language Models in Robotic Scenescited by 19
  19. 2025HSSBench: Benchmarking Humanities and Social Sciences Ability for Multimodal Large Language Modelscited by 17
  20. 2026Multimodal Multi-Agent Empowered Legal Judgment Predictioncited by 14
  21. 2026NeuReasoner: Towards Explainable, Controllable, and Unified Reasoning via Mixture-of-Neuronscited by 13
  22. 2024CODIS: Benchmarking Context-Dependent Visual Comprehension for Multimodal Large Language Modelscited by 11
  23. 2026EDIS: Diagnosing LLM Reasoning via Entropy Dynamicscited by 10
  24. 2026How Order-Sensitive Are LLMs? OrderProbe for Deterministic Structural Reconstructioncited by 9
  25. 2026FoE: Forest of Errors Makes the First Solution the Best in Large Reasoning Modelscited by 8
  26. 2026FOREVER: Forgetting Curve-Inspired Memory Replay for Language Model Continual Learningcited by 7
  27. 2026SPADER: Step-wise Peer Advantage with Diversity-Aware Exploration Rewards for Multi-Answer Question Answeringcited by 6
  28. 2026Conflict-Aware Client Selection for Multi-Server Federated Learningcited by 5
  29. 2025MUCAR: Benchmarking Multilingual Cross-Modal Ambiguity Resolution for Multimodal Large Language Modelscited by 4
  30. 2025ReasonAct: Progressive Training for Fine-Grained Video Reasoning in Small Modelscited by 4
  31. 2025Beyond N-grams: A Hierarchical Reward Learning Framework for Clinically-Aware Medical Report Generationcited by 4
  32. 2026Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial Perturbationscited by 3
  33. 2026Group Cognition Learning: Making Everything Better Through Governed Two-Stage Agents Collaborationcited by 3
  34. 2025Mitigating Multimodal Hallucinations via Gradient-based Self-Reflectioncited by 2
  35. 2026When Helpers Become Hazards: A Benchmark for Analyzing Multimodal LLM-Powered Safety in Daily Lifecited by 2
  36. 2026When Seeing Is Not Believing -- A Benchmark for Search-Grounded Video Misinformation Detectioncited by 2
  37. 2026COMET: Contrastive Motion-Enhanced Temporal Reasoning for Video Multimodal Large Language Modelscited by 1
  38. 2026Adaptive Hierarchical Representation Alliance for Multimodal Learningcited by 1

50 most cited of 53 All 53 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026HalluPrism: When Multimodal Uncertainty Should Diagnose, Not Decide
  2. 2026See It, Say It, Sorted: An Iterative Training-Free Framework for Visually-Grounded Multimodal Reasoning in LVLMs
  3. 2026Grounded Post-Training with Hard Examples for Reducing Hallucination in Multimodal Large Language Models
  4. 2026Look Carefully: Adaptive Visual Reinforcements in Multimodal Large Language Models for Hallucination Mitigation
  5. 2024HaloQuest: A Visual Hallucination Dataset for Advancing Multimodal Reasoning
  6. 2025MagiC: Evaluating Multimodal Cognition Toward Grounded Visual Reasoning
  7. 2025Socratic Questioning: Learn to Self-guide Multimodal Reasoning in the Wild
  8. 2025Measuring Epistemic Humility in Multimodal Large Language Models