mlsanityData to 14 Oct 2026

Sherlock: Self-Correcting Reasoning in Vision-Language Models

arXiv 2505.22651 ↗First posted 28 May 2025cs.CV cs.CL cs.LGcited by 13

On the map: LLM reasoning → Multimodal reasoning

also in 1 more
Abstract (arXiv)

Reasoning Vision-Language Models (VLMs) have shown promising performance on complex multimodal tasks. However, they still face significant challenges: they are highly sensitive to reasoning errors, require large volumes of annotated data or accurate verifiers, and struggle to generalize beyond specific domains. To address these limitations, we explore self-correction as a strategy to enhance reasoning VLMs. We first conduct an in-depth analysis of reasoning VLMs' self-correction abilities and identify key gaps. Based on our findings, we introduce Sherlock, a self-correction and self-improvement training framework. Sherlock introduces a trajectory-level self-correction objective, a preference data construction method based on visual perturbation, and a dynamic β for preference tuning. Once the model acquires self-correction capabilities using only 20k randomly sampled annotated data, it continues to self-improve without external supervision. Built on the Llama3.2-Vision-11B model, Sherlock achieves remarkable results across eight benchmarks, reaching an average accuracy of 64.1 with direct generation and 65.4 after self-correction. It outperforms LLaVA-CoT (63.2), Mulberry (63.9), and LlamaV-o1 (63.4) while using less than 20% of the annotated data.

Key facts. Cited by 13 papers in the map (data through 14 Oct 2026). Placed in Multimodal reasoning, one of 17 directions in LLM reasoning. Builds on 42 of its 59 references that are in the map. First posted on arXiv 28 May 2025.

Builds on

42 of 59 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025NoisyRollout: Reinforcing Visual Reasoning with Data Augmentation17 Apr
  2. 2025VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning10 Apr
  3. 2025SoTA with Less: MCTS-Guided Sample Selection for Data-Efficient Visual Reasoning Self-Improvement10 Apr
  4. 2025Skywork R1V: Pioneering Multimodal Reasoning with Chain-of-Thought8 Apr
  5. 2025Reason-RFT: Reinforcement Fine-Tuning for Visual Reasoning of Vision Language Models26 Mar
  6. 2025R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization13 Mar
  7. 2025LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL10 Mar
  8. 2025MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning10 Mar
  9. 2025MM-Verify: Enhancing Multimodal Reasoning with Chain-of-Thought Verification19 Feb
  10. 2025LlamaV-o1: Rethinking Step-by-step Visual Reasoning in LLMs10 Jan
  11. 2024MAmmoTH-VL: Eliciting Multimodal Reasoning with Instruction Tuning at Scale6 Dec
  12. 2024Critic-V: VLM Critics Help Catch VLM Errors in Multimodal Reasoning27 Nov
4 more
  1. 2024Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models21 Nov
  2. 2024LLaVA-CoT: Let Vision Language Models Reason Step-by-Step15 Nov
  3. 2024Improve Vision Language Model Chain-of-thought Reasoning21 Oct
  4. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts3 Oct

Other references · 26

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  3. 2025Qwen2.5-VL Technical Reportcited by 5,587
  4. 2024Qwen2.5 Technical Reportcited by 4,942
  5. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  6. 2024OpenAI o1 System Cardcited by 1,244
  7. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,239
  8. 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
  9. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 914
  10. 2023Gemini: A Family of Highly Capable Multimodal Modelscited by 872
  11. 2024Training Language Models to Self-Correct via Reinforcement Learningcited by 426
  12. 2025R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimizationcited by 301
14 more
  1. 2023Nash Learning from Human Feedbackcited by 224
  2. 2024Mulberry: Empowering MLLM with o1-like Reasoning and Reflection via Collective Monte Carlo Tree Searchcited by 175
  3. 2024Do Large Language Models Latently Perform Multi-Hop Reasoning?cited by 169
  4. 2024LLM Critics Help Catch LLM Bugscited by 147
  5. 2024LLaVA-Critic: Learning to Evaluate Multimodal Modelscited by 134
  6. 2025The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1cited by 91
  7. 2024Small Language Models Need Strong Verifiers to Self-Correct Reasoningcited by 86
  8. 2023LLMs cannot find reasoning errors, but can correct them given the error locationcited by 85
  9. 2025Self-rewarding correction for mathematical reasoningcited by 70
  10. 2025Critique Fine-Tuning: Learning to Critique is More Effective than Learning to Imitatecited by 63
  11. 2025Backdoor Cleaning without External Guidance in MLLM Fine-tuningcited by 35
  12. 2025Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Modelscited by 21
  13. 2024Self-Improving Robust Preference Optimizationcited by 19
  14. 2025Evolving LLMs' Self-Refinement Capability via Synergistic Training-Inference Optimizationcited by 16

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 4; 2026: 9.

13 papers.

Sort:

By area:

  1. 2025Agent0-VL: Exploring Self-Evolving Agent for Tool-Integrated Vision-Language Reasoningcited by 31
  2. 2025Generative Universal Verifier as Multimodal Meta-Reasonercited by 17
  3. 2026Does On-Policy Distillation Really Distill? From Noisy Teacher to Self-Improvementcited by 11
  4. 2026Towards Explainable Industrial Anomaly Detection via Knowledge-Guided Latent Reasoningcited by 6
  5. 2025Toward More Reliable Artificial Intelligence: Reducing Hallucinations in Vision-Language Modelscited by 1
  6. 2026Benchmarking and Evolving Reason-Reflect-Rectify for Reflective Visual Generationcited by 1
  7. 2026SVR-R1: Bootstrapping Multi-modal Reasoning with Self-verification in Reinforcement Learningcited by 1
  8. 2025Counteracting Matthew Effect in Self-Improvement of LVLMs through Head-Tail Re-balancingcited by 0
  9. 2026A Survey of Multimodal Mathematical Reasoning: From Perception, Alignment to Reasoningcited by 0
  10. 2026A Nash Equilibrium Framework For Training-Free Multimodal Step Verificationcited by 0
  11. 2026VERDICT: Training-Free Step-Wise Verification of Multimodal Reasoning via Disagreement-Aware Consensuscited by 0
  12. 2026When to Rethink: Learning Multi-Perspective Self-Verification for Vision-Language Modelscited by 0
1 more
  1. 2026The Value of Being Wrong: Self-Mined Visual In-Context Learning from Errorscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Self-Correction is More than Refinement: A Learning Framework for Visual and Language Reasoning Tasks
  2. 2024Vision-Language Models Can Self-Improve Reasoning via Reflection
  3. 2026Learn to Think: Improving Multimodal Reasoning through Vision-Aware Self-Improvement Training
  4. 2026Learning Self-Correction in Vision-Language Models via Rollout Augmentation
  5. 2025See, Think, Learn: A Self-Taught Multimodal Reasoner
  6. 2024VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning
  7. 2025VReST: Enhancing Reasoning in Large Vision-Language Models through Tree Search and Self-Reward Mechanism
  8. 2026Through the Lens of Contrast: Self-Improving Visual Reasoning in VLMs