Overthinking in Language Models

Ask about this topic

Overview

This field studies why language models sometimes keep reasoning after reaching a useful answer, repeating steps, exploring unproductive branches, or even replacing a correct answer with a wrong one. Such behavior wastes tokens and latency in tasks like math, coding, and science question answering, and can also make agent behavior vulnerable to prompts or tool calls that trigger loops. Approaches include early stopping based on confidence, answer stability, or reasoning progress; difficulty-aware compute budgets; and training or inference-time methods that prune redundant steps or steer the model out of loops. The central tension is that useful deliberation and self-correction can look like waste, so reducing length must not sacrifice accuracy, faithfulness, or the ability to handle hard problems.

Latest papers

CardsList

68 papers8 papers in the last four weeks, up 167% on the four weeks before. 0.0% of all new papers.

  1. CALM: Consistency-Aware Logit Masking in Decoding Space for Mitigating Overthinking of Large Reasoning Models

    Oct 19, 2026Shengjing Liu, Feiyang Li, Yuxuan Yang +2Overthinking in Language ModelsEfficient Language Model Reasoning

    College of Computer Science and Software Engineering, Shenzhen University

  2. Learning to Compute Is Not Learning to Retain or Reuse in Looped Transformers

    Oct 14, 2026Peilong Zhou, Hongtao Zhang, Yulong Huang +3Recurrent TransformersNeural Algorithmic Reasoning

    SKLP, Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy of Sciences · School of Advanced Interdisciplinary Sciences +1

  3. Efficiently Learning To Reason or Not to Reason: Root-token Policy Optimization for Adaptive Thinking

    Oct 13, 2026Taehyeon Kim, Hyunsoo Lee, Youngsoo Jang +1Adaptive Reasoning in Language ModelsEfficient Language Model Reasoning

    LG AI Research · UNIST · University of Illinois Chicago

  4. Constructing Pareto Frontiers of Optimal Reasoning Termination Methods for Overthinking in Reasoning Models

    Oct 9, 2026Weixin Guan, Liang Li, Shiyu Xie +7Overthinking in Language ModelsLLM Reasoning

    Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China · School of Cyberspace Security, University of Chinese Academy of Sciences, Beijing, China · University of Florida, Gainesville, FL, USA +2

  5. RATIO: Reasoning Analysis and Token-level Inference Optimization for Quantized Reasoning Models

    Sep 30, 2026Chengzhu Bao, Xianglong Yan, Tianao Zhang +3LLM QuantizationOverthinking in Language Models

    Shanghai Jiao Tong University

  6. Settle: Learning When to Stop Reasoning

    Sep 30, 2026Ryan Brown, Zihao Fu, Chris RussellEarly Stopping for Language Model ReasoningEfficient Language Model Reasoning

    Oxford Internet Institute, University of Oxford · Department of Linguistics and Modern Languages, The Chinese University of Hong Kong

  7. When Reasoning Goes Astray: Attention Dynamics of Uncontrolled Reasoning

    Sep 30, 2026Yuanhe Zhang, Ziwei Wang, Jie Ren +6Language Model SteeringOverthinking in Language Models

    Beijing University of Posts and Telecommunications · Wuhan University · Nanyang Technological University +1

  8. Efficient Reasoning via Constrained Optimization in Latent Space

    Sep 28, 2026Zhinan Hou, Xingchen Li, Keyou YouOverthinking in Language ModelsEfficient Language Model Reasoning

    Tsinghua University

  9. LLM Parkinsonism: Executive-Control Failure, Token-Inefficient Persistence, and an Uncertainty-Aware Global Executive Control Architecture for Autonomous Language-Model Agents

    Sep 25, 2026Dongsheng Xiao, Zeyuan Wang, Xuzhe Xia +2Overthinking in Language ModelsLLM Agents

    Queensland Brain Institute, The University of Queensland, Brisbane, Queensland, Australia · School of Cyber Security, Northwestern Polytechnical University, Xi'an, Shaanxi, China · Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, British Columbia, Canada +2

  10. When2Think: Learning When and How Much to Reason

    Sep 17, 2026Jaejun Shim, HyunJin Kim, Young Jin Kim +1Overthinking in Language ModelsRL for Language Model Reasoning

    Sungkyunkwan University · Microsoft

  11. SKIP: a Self-knowledge-guided Step-wise Preference Learning Framework for Concise Reasoning

    Sep 15, 2026Qinhong Lin, Yuhao Zhang, Yinglun Feng +2Efficient Language Model ReasoningLLM Reasoning

    School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing, China

  12. The Halt Vector: Internalizing a Causal Steering Intervention for Efficient Reasoning

    Aug 28, 2026Dylan Jayabahu, Tinuade AdelekeEfficient Language Model ReasoningOverthinking in Language Models

    University of Waterloo · Independent Researcher

  13. Don't Overthink, Don't Underthink: Toward Adaptive Reasoning in Agentic AI

    Aug 26, 2026Md Jueal Mia, M. Hadi AminiAgentic ReasoningOverthinking in Language Models

    Knight Foundation School of Computing and Information Sciences, Security, Optimization, and Learning for InterDependent Networks Laboratory (solid lab) Florida International University Miami, Florida, USA

  14. ChainPrune: Evaluating and Reducing Redundancy in Long Chain-of-Thought Reasoning

    Aug 22, 2026Weihang Pan, Zhengxu Yu, Yuxiang Zhang +5Efficient Language Model ReasoningOverthinking in Language Models

    Zhejiang University · Alibaba Group

  15. Fewer Tokens, Smaller Cache: Reward-Coordinated Efficient Reasoning

    Aug 5, 2026Qiyuan Zhu, Dezhi Li, Pengyu Cheng +8Overthinking in Language ModelsKV Caching

    1HKUST · 2XJTU · 3TSE +1

  16. Shorter Reasoning, Earlier Answers? An Evaluation of Reasoning Interfaces

    Aug 4, 2026Francesca Carlon, Vincent Ginis, Andres AlgabaLLM EvaluationOverthinking in Language Models

    Data Analytics Lab, Vrije Universiteit Brussel, Pleinlaan 5, 1050 Brussels, Belgium · imec-SMIT, Vrije Universiteit Brussel, Pleinlaan 9, 1050 Brussels, Belgium · School of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts 02138, USA

  17. Same Task, Different Work: Prompt-Induced Waste in Coding Agents

    Aug 2, 2026Sarel Weinberger, Amir HozezLLM Inference EfficiencyOverthinking in Language Models

    PointFive

  18. Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning

    Jul 31, 2026Xinyan Guan, Jiali Zeng, Chunlei Xin +5Overthinking in Language ModelsRL for Language Model Reasoning

    1Chinese Information Processing Laboratory, Institute of Software, Chinese Academy of Sciences · University of Chinese Academy of Sciences · 3Weixin AI, Tencent Inc, China

  19. QLPO: Quadrant-weighted Sampling for Length-aware Policy Optimization

    Jul 23, 2026Siwei Chen, Siqi Chen, Xupeng Miao +1Overthinking in Language ModelsRL for Language Model Reasoning

    School of Computer Science, Peking University · Beijing Key Laboratory of Software and Hardware Cooperative Artificial Intelligence Systems · Department of Electronic Engineering, Tsinghua University +1

  20. EvoThink: Evolving Thinking in Large Reasoning Models via Self-Pruning and Aha-Moment Preference Optimization

    Jul 22, 2026Xinbang Dai, Zheyu Xin, Huikang Hu +7Overthinking in Language ModelsSelf-Training for Language Models

    Southeast University · Noah’s Ark Lab

  21. Can We Break LLMs Out of Self-Loops? Fine-Grained Reasoning Control with Activation Steering

    Jul 20, 2026Sheldon Yu, Tong Yu, Xunyi Jiang +6Overthinking in Language ModelsCausal Interventions in Language Models

    UC San Diego · Adobe Research · University of New South Wales

  22. UPAIR: Diagnosing Reasoning States via Uncertainty-Progress Alignment for Selective Intervention

    Jul 19, 2026Cheng Yan, Zhijun Fan, Guangyang Ye +4Adaptive InferenceOverthinking in Language Models

    University of Science and Technology of China · AWS Agentic AI

  23. Valid ≠\ne Necessary: Diagnosing Latent Inefficiency in Chain-of-Thought

    Jul 13, 2026Daeyeop Lee, Hwanjo YuLLM Inference EfficiencyOverthinking in Language Models

    1KT Corporation · 2Pohang University of Science and Technology

  24. OS-Pruner: Pruning Chains-of-Thought of Reasoning Models via Optimal Stopping

    Jul 13, 2026Mohammed Ehab, Aymane El Gadarri, Vivek F. Farias +2Efficient Language Model ReasoningSequential Decision Making

    MIT · Columbia Business School

  25. Overthinking: Amplifying Reasoning Weights to Extract Learned Secrets

    Jul 9, 2026Jack Hopkins, Dipika Khullar, Fabien RogerOverthinking in Language ModelsPrivacy Leakage in Language Models

    Anthropic Fellows Program · MATS · Anthropic.

  26. Length Penalties Make Chain-of-Thought Less Monitorable

    Jul 8, 2026Bryce LittleCoT FaithfulnessOverthinking in Language Models

  27. Overthink-Triggered Slowdown Attacks on LVLM-Based Robotic Systems

    Jul 1, 2026Qiang Han, Jie Wu, Bo ChenAdversarial Attacks on VLMsOverthinking in Language Models

    Department of Computer Science, Michigan Technological University

  28. CAT: Confidence-Adaptive Thinking for Efficient Reasoning of Large Reasoning Models

    Jul 1, 2026Qizhi Jiang, Shuo Wang, Pei Ke +2Overthinking in Language ModelsEfficient Language Model Inference

    Laboratory of Intelligent Collaborative Computing, University of Electronic Science and Technology of China, Chengdu, China · Ubiquitous Intelligence and Trusted Services Key Laboratory of Sichuan Province

  29. Know When to Stop: Segment-Level Credit Assignment for Reducing Overthinking

    Jul 1, 2026Chia-Hsuan Lee, Sihui Dai, Mingyang Zhou +4Reasoning Trace AnalysisOverthinking in Language Models

    Capital One

  30. Dynamic Rollout Editing for Reducing Overthinking in RL-Trained Reasoning Models

    Jun 16, 2026Zihao Wei, Wenjie Shi, Liang Pang +8Overthinking in Language ModelsRL for Language Model Reasoning

    Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China · University of Chinese Academy of Sciences, Beijing, China