Metacognition in Language Models

Ask about this topic

Overview

The field asks whether language models can tell what they know, detect errors or changes in their internal processing, and use such assessments to decide when to revise, abstain, retrieve information, or spend more computation. Studies measure confidence calibration, error detection, self-prediction, and tool-use decisions on question-answering, reasoning, and agent benchmarks; methods include verbal confidence and token probabilities, hidden-state probes and causal interventions, and reinforcement-trained controllers. Structured reflection and adaptive control can improve accuracy or efficiency in some settings, but confidence and self-reports often fail to track correctness or transfer reliably. A central open tension is whether models have usable access to their internal states, or whether apparent metacognition reflects surface cues, task shortcuts, or external scaffolding.

Latest papers

CardsList

128 papers18 papers in the last four weeks, up 50% on the four weeks before. 0.1% of all new papers.

  1. Metacognitive Decoding: Efficient Inference for Diffusion Language Agents

    Oct 27, 2026Mingyu Zhao, Shuhang Lin, Ligong Han +5Diffusion Language Model InferenceDiffusion Language Model Decoding

    Rutgers · Amazon · UNC

  2. LADDER: Diagnosing Domain-Level Self-Assessment Under Noisy Feedback

    Oct 26, 2026Zhenshuai Yin, Jianshu Liang, Xuyue Xiong +6Language Model Self-AssessmentMetacognition in Language Models

    School of Science and Engineering, The Chinese University of Hong Kong (Shenzhen), Longgang, Shenzhen, Guangdong, 518172, P.R. China · ByteDance · Shenzhen Research Institute of Big Data, Shenzhen, China

  3. Same Beliefs, Different Policies: Learning Reusable Representations for Early Stopping in LLM Reasoning

    Oct 21, 2026Jingwei Ni, Dmitry Abulkhanov, Vladislav Smirnov +5Early Stopping for Language Model ReasoningEarly Stopping

    ETH Zürich · MBZUAI · University of Zürich +1

  4. Self-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models

    Oct 21, 2026Taewhoo Lee, Minju Song, Jungwoo Park +6RL for RAGRetrieval-Augmented Reasoning

    Korea University · Amazon · AIGEN Sciences +2

  5. Beyond Calibration: Temporal LLM Confidence Assessment Across Languages and Local Settings

    Oct 20, 2026Sahil KaleConfidence Estimation in Language ModelsMetacognition in Language Models

    University of California, Los Angeles CA, USA

  6. SOMA: Continuous Monitoring and Intervention for System Two Reasoning with System One Models

    Oct 19, 2026Ruizhi Qian, Lei Shi, Tianyu Hu +3Adaptive Reasoning in Language ModelsMetacognition in Language Models

    University of Central Florida

  7. Cross-Model In-Context Neurofeedback Reveals (Limited) Activation Control in LLMs

    Oct 14, 2026Davide Baldelli, Robyn Kozierok, Mirko Bronzi +1Metacognition in Language Models

    MATS Research · Mila – The Quebec AI Institute · LawZero

  8. A Sense of Proprioception in Language Models

    Oct 12, 2026Jonas GeipingLanguage Model IntrospectionMetacognition in Language Models

    ELLIS Institute Tübingen · Max Planck Institute for Intelligent Systems · Tübingen AI Center

  9. Collaborative Theory of Mind: Learning to Self-Revise by Anticipating User Responses

    Oct 9, 2026Fanzhi Zeng, Anubrata Das, Maytal Saar-Tsechansky +1Human-AI CollaborationTheory of Mind

    The University of Texas at Austin · Autodesk

  10. Deception by Omission: Language Models Knowingly Hide Their Mistakes

    Oct 8, 2026Lucas Florin, Amelie Knecht, Ulysse Schaller +1Deception in Language ModelsLanguage Model Error Detection

    AI Safety Research Group, University of Stuttgart, Germany

  11. When Should an In-Context Learner Expand Its Hypothesis Space?

    Oct 7, 2026Weihan Li, Xinlei Chen, Junhao Wu +1In-Context LearningMetacognition in Language Models

    Department of Mechanical Engineering, The University of Tokyo, Tokyo, Japan · School of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen, China · Faculty of Mechanical Engineering, RWTH Aachen University, Aachen, Germany +1

  12. Test-Time Adaptation of Reasoning Strategies with Bayesian Nonparametric Memory

    Oct 5, 2026Keshav Ramji, Tahira Naseem, Ramón Fernandez AstudilloMemory-Augmented Language ModelsBayesian Mixture Models

    IBM Research AI

  13. Anosognosia in LLMs: Probing Self-Awareness of Quantized Computational Substrate

    Oct 5, 2026Yoshihiro Izawa, Gouki Minegishi, Yoko YamakataLanguage Model Self-AssessmentMetacognition in Language Models

    The University of Tokyo

  14. Look Before You Leap: Thermodynamic Arbitration of Parametric and Non-Parametric Knowledge in LLM Agents via Self-Regulating Memory Architectures

    Oct 4, 2026Akash Das, Ishan RoyCognitive Architectures for AI AgentsAdaptive RAG

    Fidelity Investments

  15. Agentic Cognitive Depth: Operational Criteria for Evaluating LLM Agents

    Oct 3, 2026Nijesh Upreti, Chris Sypherd, Vaishak BelleLLM Agent EvaluationLLM Agents

    The University of Edinburgh 10 Crichton Street, Edinburgh EH8 9AB, UK

  16. MetaCtrl: Your Large Language Models Can Reason Better and More Concisely with a Metacognitive Controller

    Sep 29, 2026Zhibin Wen, Tao Han, Lei Bai +2Adaptive Reasoning in Language ModelsRL for Language Model Reasoning

    Southern University of Science and Technology · The Hong Kong University of Science and Technology · Shanghai Artificial Intelligence Laboratory +1

  17. AwarenessBench: Assessing Cognitive Capabilities of Language Models

    Sep 28, 2026Xiaojian Li, Rongwu Xu, Tianyun Zhang +9LLM EvaluationEvaluation Awareness in Language Models

    Tsinghua University · Shanghai Qi Zhi Institute · Fangcun AI +5

  18. A mechanistic study of language model introspection

    Sep 28, 2026Jiahong Zou, Xiangkun Sun, Lingkai Kong +1Attention Head AnalysisLLM Interpretability

    Shandong University · Tsinghua University · Northeastern University +1

  19. On the Limits of Metacognitive Monitoring in LLMs

    Sep 28, 2026Dongqi Han, Yifan Yang, Dongsheng LiLanguage Model Error DetectionConfidence Estimation in Language Models

    Microsoft Research Asia

  20. Nudgeability: Reasoning Models Follow Confidence Signals Without Tracking Their Own Competence

    Sep 28, 2026Rohit Saxena, Utkarsh UpadhyayLanguage Model Self-AssessmentLLM Tool Use

  21. LLMs learn different forms of metacognition when trained to predict their own accuracy

    Sep 27, 2026Nicolas Yax, Stefano Palminteri, Pierre-Yves OudeyerConfidence Estimation in Language ModelsLanguage Model Calibration

    LNC2, INSERM, Paris, France · DEC, ENS, PSL, Paris, France · Flowers AI & CogSci Lab +1

  22. Self-Reports Do Not Identify Self-Models: An Identifiability Test for Counterfactual Reports

    Sep 26, 2026Phongsakon Mark Konrad, Toygar Tanyel, Serkan AyvazLanguage Model Self-AssessmentMetacognition in Language Models

    Centre for Industrial Software, University of South- ern Denmark, Sønderborg, Denmark · Promake, Newark, DE, USA.

  23. Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

    Sep 25, 2026Parsa Hosseini, Akasha Tigalappanavara, Sumit Nawathe +6Efficient Language Model ReasoningMetacognition in Language Models

    University of Maryland · AI Foundations, Capital One

  24. Confidence Comes from Experience: Experiential Confidence Estimation from Reasoning to Agents

    Sep 15, 2026Caiqi Zhang, Xiaochen Zhu, Chengzu Li +3Confidence Estimation in Language ModelsLLM Uncertainty Estimation

    University of Cambridge 1 Solve & Estimate · Google DeepMind

  25. Self-Emergence Agent Architecture:Behavior-Inertia HMM, Reflexive Metacognition,and Social-Contrastive Self-Modeling

    Sep 15, 2026Xiaoyang LiuSelf-Evolving AgentsCognitive Architectures for AI Agents

    Independent Researcher

  26. Metacognitive Steering: Learning the Structure of Scientific Judgment

    Sep 14, 2026Vincent Karpf, Joseph Reth, Eike Gerhardt +5Inference-Time GuidanceScientific Reasoning in Language Models

    Autopoiesis Sciences

  27. Rethinking Verbalized Confidence for LLM-as-a-Judge: A Compatibility Shift on Post-2025 Proprietary Models

    Sep 10, 2026Yu-Chung HsiaoLanguage Model CalibrationLLM-as-a-Judge

    Cisco Systems

  28. Strangers to Themselves: What Language Models Say About Themselves Is Generic

    Sep 9, 2026Phil Blandfort, Urja PawarLanguage Model Self-AssessmentMetacognition in Language Models

    Predictably Weird · Independent

  29. "very likely" Means "uncertain"? How LLMs Diverge from Humans in Linguistic Uncertainty Quantification

    Sep 6, 2026Jinhao Duan, Zicheng Liu, Zijie Liu +2LLM Uncertainty EstimationVerbalized Confidence Estimation

    Department of Computer Science, UNC-Chapel Hill · Department of Mathematics, The University of Hong Kong · Department of Data Science, City University of Hong Kong

  30. A Calibrated Reflection Approach for Enhancing Confidence Estimation in LLMs

    Sep 3, 2026Umesh Bodhwani, Yuan Ling, Shujing Dong +3Confidence Estimation in Language ModelsLanguage Model Calibration

    Amazon.com Seattle, Washington, USA