mlsanityData to 7 Oct 2026

Confidence Matters: Revisiting Intrinsic Self-Correction Capabilities of Large Language Models

arXiv 2402.12563 ↗First posted 19 Feb 2024cs.CL cs.AIcited by 47

Not on the map. Why?

Abstract (arXiv)

The recent success of Large Language Models (LLMs) has catalyzed an increasing interest in their self-correction capabilities. This paper presents a comprehensive investigation into the intrinsic self-correction of LLMs, attempting to address the ongoing debate about its feasibility. Our research has identified an important latent factor - the "confidence" of LLMs - during the self-correction process. Overlooking this factor may cause the models to over-criticize themselves, resulting in unreliable conclusions regarding the efficacy of self-correction. We have experimentally observed that LLMs possess the capability to understand the "confidence" in their own responses. It motivates us to develop an "If-or-Else" (IoE) prompting framework, designed to guide LLMs in assessing their own "confidence", facilitating intrinsic self-corrections. We conduct extensive experiments and demonstrate that our IoE-based Prompt can achieve a consistent improvement regarding the accuracy of self-corrected responses over the initial answers. Our study not only sheds light on the underlying factors affecting self-correction in LLMs, but also introduces a practical framework that utilizes the IoE prompting principle to efficiently improve self-correction capabilities with "confidence". The code is available at https://github.com/MBZUAI-CLeaR/IoE-Prompting.git.

Key facts. Cited by 47 papers in the map (data through 7 Oct 2026). Builds on 45 of its 60 references that are in the map. First posted on arXiv 19 Feb 2024.

Builds on

45 of 60 references.

References · 45

  1. 2023GPT-4 Technical Reportcited by 18,132
  2. 2022Training language models to follow instructions with human feedbackcited by 15,712
  3. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  4. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  5. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,686
  6. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,634
  7. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
  8. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  9. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,970
  10. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  11. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,739
  12. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 965
33 more
  1. 2023Teaching Large Language Models to Self-Debugcited by 963
  2. 2021Are NLP Models really able to Solve Simple Math Word Problems?cited by 861
  3. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 751
  4. 2022Improving alignment of dialogue agents via targeted human judgementscited by 527
  5. 2023Fine-Grained Human Feedback Gives Better Rewards for Language Model Trainingcited by 399
  6. 2023Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoningcited by 379
  7. 2022Self-critiquing models for assisting human evaluatorscited by 372
  8. 2023Language Models can Solve Computer Taskscited by 371
  9. 2023Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedbackcited by 351
  10. 2022RARR: Researching and Revising What Language Models Say, Using Language Modelscited by 286
  11. 2022Generating Sequences by Learning to Self-Correctcited by 271
  12. 2023FacTool: Factuality Detection in Generative AI -- A Tool Augmented Framework for Multi-Task and Multi-Domain Scenarioscited by 247
  13. 2023LLM Self Defense: By Self Examination, LLMs Know They Are Being Trickedcited by 226
  14. 2023REFINER: Reasoning Feedback on Intermediate Representationscited by 203
  15. 2023Improving Language Model Negotiation with Self-Play and In-Context Learning from AI Feedbackcited by 202
  16. 2022Re3: Generating Longer Stories With Recursive Reprompting and Revisioncited by 193
  17. 2023Is Self-Repair a Silver Bullet for Code Generation?cited by 191
  18. 2023The Capacity for Moral Self-Correction in Large Language Modelscited by 171
  19. 2023Rephrase and Respond: Let Large Language Models Ask Better Questions for Themselvescited by 123
  20. 2023PRD: Peer Rank and Discussion Improve Large Language Model based Evaluationscited by 114
  21. 2023Self-Edit: Fault-Aware Code Editor for Code Generationcited by 110
  22. 2023RL4F: Generating Natural Language Feedback with Reinforcement Learning for Repairing Model Outputscited by 104
  23. 2023GPT-4 Doesn't Know It's Wrong: An Analysis of Iterative Prompting for Reasoning Problemscited by 88
  24. 2023Improving Language Models via Plug-and-Play Retrieval Feedbackcited by 87
  25. 2023LLMs cannot find reasoning errors, but can correct them given the error locationcited by 85
  26. 2023FLIRT: Feedback Loop In-context Red Teamingcited by 81
  27. 2022Controllable Natural Language Generation with Contrastive Prefixescited by 75
  28. 2023Self-Verification Improves Few-Shot Clinical Information Extractioncited by 52
  29. 2023ALGO: Synthesizing Algorithmic Programs with LLM-Generated Oracle Verifierscited by 41
  30. 2023BenchLMM: Benchmarking Cross-style Visual Capability of Large Multimodal Modelscited by 41
  31. 2023Teaching Language Models to Self-Improve through Interactive Demonstrationscited by 26
  32. 2023Learning to Simulate Natural Language Feedback for Interactive Semantic Parsingcited by 19
  33. 2022Instance-aware Prompt Learning for Language Understanding and Generationcited by 18

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 13; 2025: 20; 2026: 14.

47 papers.

Sort:

By area:

  1. 2024When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMscited by 230
  2. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Surveycited by 195
  3. 2025A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligencecited by 155
  4. 2024On Verbalized Confidence Scores for LLMscited by 83
  5. 2025Lifelong Learning of Large Language Model based Agents: A Roadmapcited by 78
  6. 2024Latent Space Chain-of-Embedding Enables Output-free LLM Self-Evaluationcited by 56
  7. 2024A Theoretical Understanding of Self-Correction through In-context Alignmentcited by 43
  8. 2024Self-Correction is More than Refinement: A Learning Framework for Visual and Language Reasoning Taskscited by 30
  9. 2025Low-probability Tokens Sustain Exploration in Reinforcement Learning with Verifiable Rewardcited by 25
  10. 2024Causality for Large Language Modelscited by 24
  11. 2024Reinforcement Learning from Multi-role Debates as Feedback for Bias Mitigation in LLMscited by 22
  12. 2024Can Large Vision-Language Models Correct Semantic Grounding Errors By Themselves?cited by 20
35 more
  1. 2024On the Intrinsic Self-Correction Capability of LLMs: Uncertainty and Latent Conceptcited by 16
  2. 2025Generative AI Act II: Test Time Scaling Drives Cognition Engineeringcited by 14
  3. 2026Not All Queries Need Deep Thought: CoFiCot for Adaptive Coarse-to-fine Stateful Refinementcited by 13
  4. 2025Meta-Thinking in LLMs via Multi-Agent Reinforcement Learning: A Surveycited by 10
  5. 2024A Closer Look at Logical Reasoning with LLMs: The Choice of Tool Matterscited by 8
  6. 2025Beyond Pipelines: A Survey of the Paradigm Shift toward Model-Native Agentic AIcited by 8
  7. 2026Revision or Re-Solving? Decomposing Second-Pass Gains in Multi-LLM Pipelinescited by 7
  8. 2024FIRE: A Dataset for Feedback Integration and Refinement Evaluation of Multimodal Modelscited by 6
  9. 2024Re-Thinking Process Mining in the AI-Based Agents Eracited by 5
  10. 2025Entropy-Based Adaptive Weighting for Self-Trainingcited by 5
  11. 2026Self-Improvement of Large Language Models: A Technical Overview and Future Outlookcited by 5
  12. 2025Can LLMs Detect Their Confabulations? Estimating Reliability in Uncertainty-Aware Language Modelscited by 4
  13. 2025Towards Self-Refinement of Vision-Language Models with Triangular Consistencycited by 3
  14. 2026How LLMs Detect and Correct Their Own Errors: The Role of Internal Confidence Signalscited by 3
  15. 2024Automating Intervention Discovery from Scientific Literature: A Progressive Ontology Prompting and Dual-LLM Frameworkcited by 2
  16. 2025Explaining Intrinsic Moral Self-Correction with Mechanistic Interpretabilitycited by 2
  17. 2025Adaptive and Resource-efficient Agentic AI Systems for Mobile and Embedded Devices: A Surveycited by 2
  18. 2025Beware of Reasoning Overconfidence: Pitfalls in the Reasoning Process for Multi-solution Taskscited by 2
  19. 2026Self-Debias: Self-correcting for Debiasing Large Language Modelscited by 2
  20. 2025Contemplative Artificial Intelligencecited by 1
  21. 2025RETuning: Upgrading Inference-Time Scaling for Stock Movement Prediction with Large Language Modelscited by 1
  22. 2025A Data-Centric Approach for Safe and Secure Large Language Models against Threatening and Toxic Contentcited by 0
  23. 2025HAVE: Head-Adaptive Gating and ValuE Calibration for Hallucination Mitigation in Large Language Modelscited by 0
  24. 2025Unleashing the True Potential of LLMs: A Feedback-Triggered Self-Correction with Long-Term Multipath Decodingcited by 0
  25. 2025PersonaX: Multimodal Datasets with LLM-Inferred Behavior Traitscited by 0
  26. 2025Sample, Align, Synthesize: Graph-Based Response Synthesis with ConGrscited by 0
  27. 2026Cognitive Loop of Thought: Reversible Hierarchical Markov Chain for Efficient Mathematical Reasoningcited by 0
  28. 2026When to Vote, When to Rewrite: Disagreement-Guided Strategy Routing for Test-Time Scalingcited by 0
  29. 2026No Action Without a NOD: A Heterogeneous Multi-Agent Architecture for Reliable Service Agentscited by 0
  30. 2026DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimizationcited by 0
  31. 2026Integrating Local and Global Entropy for Uncertainty Quantification in LLMscited by 0
  32. 2026Spectral Origins of the Self-Correction Blind Spot in Autoregressive Generationcited by 0
  33. 2026DC-Leap: Training-Free Acceleration of dLLMs via Draft-Guided Contiguous Leaping Decodingcited by 0
  34. 2026The Calibration Floor: Format Repair Can Masquerade as Self-Correction at Small-to-Mid Scalecited by 0
  35. 2026Beyond Solo and Consistency: Vindicating Multi-Agent Debate via Conditional Progressive Pruningcited by 0