mlsanityData to 7 Oct 2026

ReVISE: Learning to Refine at Test-Time via Intrinsic Self-Verification

arXiv 2502.14565 ↗First posted 20 Feb 2025cs.LG cs.CLcited by 35

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Self-awareness, i.e., the ability to assess and correct one's own generation, is a fundamental aspect of human intelligence, making its replication in large language models (LLMs) an important yet challenging task. Previous works tackle this by employing extensive reinforcement learning or rather relying on large external verifiers. In this work, we propose Refine via Intrinsic Self-Verification (ReVISE), an efficient and effective framework that enables LLMs to self-correct their outputs through self-verification. The core idea of ReVISE is to enable LLMs to verify their reasoning processes and continually rethink reasoning trajectories based on its verification. We introduce a structured curriculum based upon online preference learning to implement this efficiently. Specifically, as ReVISE involves two challenging tasks (i.e., self-verification and reasoning correction), we tackle each task sequentially using curriculum learning, collecting both failed and successful reasoning paths to construct preference pairs for efficient training. During inference, our approach enjoys natural test-time scaling by integrating self-verification and correction capabilities, further enhanced by our proposed confidence-aware decoding mechanism. Our experiments on various reasoning tasks demonstrate that ReVISE achieves efficient self-correction and significantly improves reasoning performance.

Key facts. Cited by 35 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 38 of its 42 references that are in the map. First posted on arXiv 20 Feb 2025.

Builds on

38 of 42 references.

Closest ancestors · 10

Same direction, newest first

  1. 2024Improving LLM Reasoning through Scaling Inference Computation with Collaborative Verification5 Oct
  2. 2024Training Language Models to Self-Correct via Reinforcement Learning19 Sep
  3. 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters6 Aug
  4. 2024V-STaR: Training Verifiers for Self-Taught Reasoners9 Feb
  5. 2024ReFT: Reasoning with Reinforced Fine-Tuning17 Jan
  6. 2023Large Language Models Cannot Self-Correct Reasoning Yet3 Oct
  7. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Models17 May
  8. 2023Self-Refine: Iterative Refinement with Self-Feedback30 Mar
  9. 2022Large Language Models Can Self-Improve20 Oct
  10. 2022STaR: Bootstrapping Reasoning With Reasoning28 Mar

Other references · 28

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2024The Llama 3 Herd of Modelscited by 14,192
  3. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  4. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  5. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
  6. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  7. 2022Large Language Models are Zero-Shot Reasonerscited by 4,323
  8. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,064
  9. 2023Let's Verify Step by Stepcited by 3,598
  10. 2021Program Synthesis with Large Language Modelscited by 3,509
  11. 2022Emergent Abilities of Large Language Modelscited by 2,483
  12. 2023MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Modelscited by 724
16 more
  1. 2024Improve Mathematical Reasoning in Language Models by Automated Process Supervisioncited by 426
  2. 2023Scaling Relationship on Learning Mathematical Reasoning with Large Language Modelscited by 417
  3. 2024CodeAgent: Enhancing Code Generation with Tool-Integrated Agent Systems for Real-World Repo-level Coding Challengescited by 278
  4. 2024Trial and Error: Exploration-Based Trajectory Optimization for LLM Agentscited by 194
  5. 2023Speculative Decoding with Big Little Decodercited by 163
  6. 2024Recursive Introspection: Teaching Language Model Agents How to Self-Improvecited by 162
  7. 2024The pitfalls of next-token predictioncited by 156
  8. 2024Accessing GPT-4 level Mathematical Olympiad Solutions via Monte Carlo Tree Self-refine with LLaMa-3 8Bcited by 156
  9. 2024Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizercited by 109
  10. 2024When Search Engine Services meet Large Language Models: Visions and Challengescited by 59
  11. 2024Backtracking Improves Generation Safetycited by 31
  12. 2024Do LLMs Exhibit Human-Like Reasoning? Evaluating Theory of Mind in LLMs for Open-Ended Responsescited by 22
  13. 2023Artificial Intelligence-Enabled Intelligent Assistant for Personalized and Adaptive Learning in Higher Educationcited by 17
  14. 2024A Human-Like Reasoning Framework for Multi-Phases Planning Task with Large Language Modelscited by 12
  15. 2024System-2 Mathematical Reasoning via Enriched Instruction Tuningcited by 7
  16. 2024System 2 Reasoning Capabilities Are Nighcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 18; 2026: 17.

35 papers.

Sort:

By area:

  1. 2025Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Modelscited by 381
  2. 2025DiMo-GUI: Advancing Test-time Scaling in GUI Grounding via Modality-Aware Visual Reasoningcited by 29
  3. 2025Thinking Machines: A Survey of LLM based Reasoning Strategiescited by 25
  4. 2025Variation in Verification: Understanding Verification Dynamics in Large Language Modelscited by 18
  5. 2025Think Clearly: Improving Reasoning via Redundant Token Pruningcited by 17
  6. 2025MARS: toward more efficient multi-agent collaboration for LLM reasoningcited by 13
  7. 2025ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Searchcited by 12
  8. 2026CoVerRL: Breaking the Consensus Trap in Label-Free Reasoning via Generator-Verifier Co-Evolutioncited by 9
  9. 2025Efficient LLM Collaboration via Planningcited by 7
  10. 2025AgentAsk: Multi-Agent Systems Need to Askcited by 6
  11. 2025Monitor-Generate-Verify (MGV): Formalising Metacognitive Theory for Language Model Reasoningcited by 4
  12. 2026Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface Elementscited by 4
23 more
  1. 2026On the Generalization Gap in Self-Evolving Language Model Reasoningcited by 4
  2. 2025Self-Correcting Code Generation Using Small Language Modelscited by 3
  3. 2025Before you <think>, monitor: Implementing Flavell's metacognitive framework in LLMscited by 3
  4. 2025Mind the Gap: Bridging Thought Leap for Improved Chain-of-Thought Tuningcited by 2
  5. 2025Scalable and Robust LLM Unlearning by Correcting Responses with Retrieved Exclusionscited by 2
  6. 2025SDA: Steering-Driven Distribution Alignment for Open LLMs without Fine-Tuningcited by 2
  7. 2026Self-Debias: Self-correcting for Debiasing Large Language Modelscited by 2
  8. 2025Test-Time Scaling Strategies for Generative Retrieval in Multimodal Conversational Recommendationscited by 1
  9. 2026Beyond In-Domain Detection: SpikeScore for Cross-Domain Hallucination Detectioncited by 1
  10. 2026Beyond Surface Imitation: Contrastive Modeling for Reasoning Path Alignment in Multimodal In-Context Learningcited by 1
  11. 2025Reasoning With a Star: A Heliophysics Dataset and Benchmark for Agentic Scientific Reasoningcited by 0
  12. 2025CADMorph: Geometry-Driven Parametric CAD Editing via a Plan-Generate-Verify Loopcited by 0
  13. 2026Value-Aware Numerical Representations for Transformer Language Modelscited by 0
  14. 2026MarkovScale: Towards Optimal Sequential Scaling at Inference Timecited by 0
  15. 2026Internalizing Outcome Supervision into Process Supervision: A New Paradigm for Reinforcement Learning for Reasoningcited by 0
  16. 2026Primal Generation, Dual Judgment: Self-Training from Test-Time Scalingcited by 0
  17. 2026SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptationcited by 0
  18. 2026REVES: REvision and VErification--Augmented Training for Test-Time Scalingcited by 0
  19. 2026Spectral Origins of the Self-Correction Blind Spot in Autoregressive Generationcited by 0
  20. 2026Refining Over Resampling: Test-Time Self-Correction for LLM Reasoningcited by 0
  21. 2026DCGC: Draft-Conditioned Global Correction for Complex Reasoning with Masked Diffusion Modelscited by 0
  22. 2026CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agentscited by 0
  23. 2026Provable Test-Time Scaling for Beam Search in LLM Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026SVR: Self-Verifying Refinement via Joint Verdict-Confidence Reinforcement Learning for Adaptive Test-Time Compute
  2. 2025PAG: Multi-Turn Reinforced LLM Self-Correction with Policy as Generative Verifier
  3. 2026Training Large Language Models for Self-Explanation Faithfulness
  4. 2026Learning to Self-Verify Makes Language Models Better Reasoners
  5. 2024Self-Refine Instruction-Tuning for Aligning Reasoning in Language Models
  6. 2026A Training-Free Regeneration Paradigm: Contrastive Reflection Memory Guided Self-Verification and Self-Improvement
  7. 2025Evolving LLMs' Self-Refinement Capability via Synergistic Training-Inference Optimization
  8. 2024Towards Intrinsic Self-Correction Enhancement in Monte Carlo Tree Search Boosted Reasoning via Iterative Preference Learning