mlsanityData to 7 Oct 2026

ClueTracer: Question-to-Vision Clue Tracing for Training-Free Hallucination Suppression in Multimodal Reasoning

arXiv 2602.02004 ↗First posted 2 Feb 2026cs.CV cs.AIcited by 0

On the map: LLM reasoning → Multimodal reasoning

Abstract (arXiv)

Large multimodal reasoning models solve challenging visual problems via explicit long-chain inference: they gather visual clues from images and decode clues into textual tokens. Yet this capability also increases hallucinations, where the model generates content that is not supported by the input image or the question. To understand this failure mode, we identify reasoning drift: during clue gathering, the model over-focuses on question-irrelevant entities, diluting focus on task-relevant cues and gradually decoupling the reasoning trace from visual grounding. As a consequence, many inference-time localization or intervention methods developed for non-reasoning models fail to pinpoint the true clues in reasoning settings. Motivated by these insights, we introduce ClueRecall, a metric for assessing visual clue retrieval, and present ClueTracer, a training-free, parameter-free, and architecture-agnostic plugin for hallucination suppression. ClueTracer starts from the question and traces how key clues propagate along the model's reasoning pathway (question rightarrow outputs rightarrow visual tokens), thereby localizing task-relevant patches while suppressing spurious attention to irrelevant regions. Remarkably, without any additional training, ClueTracer improves all reasoning architectures (including textttR1-OneVision, textttOcean-R1, textttMM-Eureka, etc.) by 1.21× on reasoning benchmarks. When transferred to non-reasoning settings, it yields a 1.14× gain.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 43 of its 64 references that are in the map. First posted on arXiv 2 Feb 2026.

Builds on

43 of 64 references.

Closest ancestors · 19

Same direction, newest first

  1. 2025CoRGI: Verified Chain-of-Thought Reasoning with Post-hoc Visual Grounding1 Aug
  2. 2025GoT-R1: Unleashing Reasoning Capability of MLLM for Visual Generation with Reinforcement Learning22 May
  3. 2025Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models22 May
  4. 2025VLM-R3: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought22 May
  5. 2025GRIT: Teaching MLLMs to Think with Images21 May
  6. 2025Adaptive Chain-of-Focus Reasoning via Dynamic Visual Search and Zooming for Efficient VLMs21 May
  7. 2025SoTA with Less: MCTS-Guided Sample Selection for Data-Efficient Visual Reasoning Self-Improvement10 Apr
  8. 2025Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning of Vision Language Models26 Mar
  9. 2025Multimodal Chain-of-Thought Reasoning: A Comprehensive Survey16 Mar
  10. 2025R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization13 Mar
  11. 2025VisualPRM: An Effective Process Reward Model for Multimodal Reasoning13 Mar
  12. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
7 more
  1. 2025Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models9 Mar
  2. 2025Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement9 Mar
  3. 2025LlamaV-o1: Rethinking Step-by-step Visual Reasoning in LLMs10 Jan
  4. 2024Improve Vision Language Model Chain-of-thought Reasoning21 Oct
  5. 2024Measuring Multimodal Mathematical Reasoning with MATH-Vision Dataset22 Feb
  6. 2023Multimodal Chain-of-Thought Reasoning in Language Models2 Feb
  7. 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering20 Sep

Other references · 24

  1. 2023Visual Instruction Tuningcited by 7,792
  2. 2025Qwen2.5-VL Technical Reportcited by 5,418
  3. 2024Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMscited by 719
  4. 2024Hallucination of Multimodal Large Language Models: A Surveycited by 380
  5. 2025R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimizationcited by 296
  6. 2024Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decodingcited by 198
  7. 2024Mulberry: Empowering MLLM with o1-like Reasoning and Reflection via Collective Monte Carlo Tree Searchcited by 171
  8. 2024Marco-o1: Towards Open Reasoning Models for Open-Ended Solutionscited by 142
  9. 2025See What You Are Told: Visual Attention Sink in Large Multimodal Modelscited by 127
  10. 2025Why Is Spatial Reasoning Hard for VLMs? An Attention Mechanism Perspective on Focus Areascited by 93
  11. 2024LLaMA-Berry: Pairwise Optimization for O1-like Olympiad-Level Mathematical Reasoningcited by 90
  12. 2025More Thinking, Less Seeing? Assessing Amplified Hallucination in Multimodal Reasoning Modelscited by 84
12 more
  1. 2024Look Twice Before You Answer: Memory-Space Visual Retracing for Hallucination Mitigation in Multimodal Large Language Modelscited by 72
  2. 2025The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models via Visual Information Steeringcited by 62
  3. 2025RedStar: Does Scaling Long-CoT Data Unlock Better Slow-Reasoning Systems?cited by 60
  4. 2024Mitigating Modality Prior-Induced Hallucinations in Multimodal Large Language Models via Deciphering Attention Causalitycited by 42
  5. 2024Visual Description Grounding Reduces Hallucinations and Boosts Reasoning in LVLMscited by 33
  6. 2025One RL to See Them All: Visual Triple Unified Reinforcement Learningcited by 29
  7. 2025Auditing Meta-Cognitive Hallucinations in Reasoning Large Language Modelscited by 19
  8. 2025MIRAGE: Assessing Hallucination in Multimodal Reasoning Chains of MLLMcited by 17
  9. 2025CLAIM: Mitigating Multilingual Object Hallucination in Large Vision-Language Models with Cross-Lingual Attention Interventioncited by 11
  10. 2025Mitigating Hallucinations in Large Vision-Language Models via Entity-Centric Multimodal Preference Optimizationcited by 10
  11. 2023Fast Chain-of-Thought: A Glance of Future from Parallel Decoding Leads to Answers Fastercited by 5
  12. 2025Self-Aug: Query and Entropy Adaptive Decoding for Large Vision-Language Modelscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Cognitive Pivot Points and Visual Anchoring: Unveiling and Rectifying Hallucinations in Multimodal Reasoning Models
  2. 2024Thinking Before Looking: Improving Multimodal LLM Reasoning via Mitigating Visual Hallucination
  3. 2026Seeing Through the Chain: Mitigate Hallucination in Multimodal Reasoning Models via CoT Compression and Contrastive Preference Optimization
  4. 2024HaloQuest: A Visual Hallucination Dataset for Advancing Multimodal Reasoning
  5. 2026See It, Say It, Sorted: An Iterative Training-Free Framework for Visually-Grounded Multimodal Reasoning in LVLMs
  6. 2026Reflect to Inform: Boosting Multimodal Reasoning via Information-Gain-Driven Verification
  7. 2025Visual Attention Reasoning via Hierarchical Search and Self-Verification
  8. 2025Draft and Refine with Visual Experts