CoT Reasoning

CoT: Chain-of-Thought

Chain-of-thought reasoning studies how language models break complex problems into steps, use intermediate information, and produce answers that can be checked or acted on. Approaches include prompting models to show their work, training on reasoning traces with supervised learning or reinforcement learning, and using structured methods such as code, symbolic rules, trees, or latent reasoning states. Researchers test these methods on tasks including mathematics, multi-hop question answering, coding, clinical decisions, and visual or spatial understanding, while developing verification and adaptive-compute methods to control cost. A central tension is that longer or more detailed traces do not reliably improve accuracy or explain how a model reached its answer; they can also amplify errors, bias, or unsupported claims.

Latest papers

228 papers25 papers in the last four weeks, up 56% on the four weeks before. 0.1% of all new papers.

Calibration Drift Under Reasoning: How Chain-of-Thought Budgets Induce Overconfidence in Large Language Models
Apr 24, 2026Prakul Sunil Hiremath, Harshit R. HiremathConfidence Estimation in Language Models CoT Reasoning
Department of Computer Science and Engineering Visvesvaraya Technological University, Belagavi · Department of Computer Science and Business System SG Balekundri Institute of Technology, Belagavi
Reasoning Primitives in Hybrid and Non-Hybrid LLMs: Do Architectural Differences Yield Advantages in State-Tracking and Recall?
Apr 23, 2026Shivam Rawat, Lucie Flek, Florian Mai +1CoT Reasoning LLM Memory
Lamarr Institute for Machine Learning and Artificial Intelligence · Rheinische Friedrich-Wilhelms-Universität Bonn
VG-CoT: Towards Trustworthy Visual Reasoning via Grounded Chain-of-Thought
Apr 23, 2026Byeonggeuk Lim, Kyeonghyun Kim, JungMin Yun +1VLM Evaluation CoT Reasoning
Graduate School of Advanced Imaging Science, Multimedia & Film, Chung-Ang University1 · Department of Artificial Intelligence, Chung-Ang University2
Less Languages, Less Tokens: An Efficient Unified Logic Cross-lingual Chain-of-Thought Reasoning Framework
Apr 22, 2026Chenyuan Zhang, Qiguang Chen, Xie Chen +6Cross-Lingual Reasoning in Language Models CoT Reasoning
Harbin Institute of Technology, Shenzhen · Shanghai Innovation Institute · Central South University +3
Does Self-Consistency Improve the Recall of Encyclopedic Knowledge?
Apr 21, 2026Sho Hoshino, Ukyo Honda, Peinan ZhangSelf-Consistency Decoding CoT Reasoning
CyberAgent
More Thinking, More Bias: Length-Driven Position Bias in Reasoning Models
Apr 21, 2026Xiao WangMultiple-Choice Question Answering CoT Reasoning
PRISMA: Preference-Reinforced Self-Training Approach for Interpretable Emotionally Intelligent Negotiation Dialogues
Apr 20, 2026Prajwal Vijay Kajare, Priyanshu Priya, Bikash Santra +1Automated Negotiation CoT Reasoning
School of Artificial Intelligence and Data Science, Indian Institute of Technology Jodhpur, India · Department of Computer Science and Engineering, Indian Institute of Technology Patna, India
ContraPrompt: Contrastive Prompt Optimization via Dyadic Reasoning Trace Analysis
Apr 20, 2026Rishav Rishav, Pushpak Pujari, Pushpendre RastogiPairwise Comparison CoT Reasoning
Vizops AI
Learning to Correct: Calibrated Reinforcement Learning for Multi-Attempt Chain-of-Thought
Apr 20, 2026Muhammed Emrullah Ildiz, Halil Alperen Gozeten, Ege Onur Taga +1CoT Reasoning RL for Language Model Reasoning
University of Michigan, Ann Arbor · Google Research
Self-Consistency from Only Two Samples: CoT-PoT Ensembling for Efficient LLM Reasoning
Apr 19, 2026Raman Saparkhan, Majd Hawasly, Md Rizwan Parvez +1LLM Inference Efficiency Self-Consistency Decoding
Carnegie Mellon University Qatar · Qatar Computing Research Institute
Clover: A Neural-Symbolic Agentic Harness with Stochastic Tree-of-Thoughts for Verified RTL Repair
Apr 19, 2026Zizhang Luo, Yansong Xu, Runlin Guo +6Automated Program Repair Tree of Thoughts
Peking University
LLM Reasoning Is Latent, Not the Chain of Thought
Apr 17, 2026Wenshuo WangLLM Interpretability CoT Reasoning
School of Future Technology, South China University of Technology, China
MemoSight: Unifying Context Compression and Multi Token Prediction for Reasoning Acceleration
Apr 16, 2026Xinyu Liu, Xin Liu, Bo Jin +8Multi-Token Prediction KV Caching
School of Computer Science and Engineering, Northeastern University, China · Meituan Inc. · NiuTrans Research, Shenyang, China
Analyzing Chain of Thought (CoT) Approaches in Control Flow Code Deobfuscation Tasks
Apr 16, 2026Seyedreza Mohseni, Sarvesh Baskar, Edward Raff +1Software Reverse Engineering CoT Reasoning
University of Maryland Baltimore County, Baltimore, USA · CrowdStrike, New York, USA
On the Expressive Power and Limitations of Multi-Layer SSMs
Apr 16, 2026Nikola Zubić, Qian Li, Yuyi Wang +1CoT Reasoning State Space Models
Robotics and Perception Group, University of Zurich · Shenzhen International Center For Industrial And Applied Mathematics, Shenzhen Research Institute of Big Data · Tengen Intelligence Institute, CRRC Zhuzhou Institute
LongCoT: Benchmarking Long-Horizon Chain-of-Thought Reasoning
Apr 15, 2026Sumeet Ramesh Motwani, Daniel Nichols, Charles London +17Long-Horizon LLM Reasoning CoT Reasoning
University of Oxford · Lawrence Livermore National Laboratory (LLNL) · MBZUAI +4
Correct Prediction, Wrong Steps? Consensus Reasoning Knowledge Graph for Robust Chain-of-Thought Synthesis
Apr 15, 2026Zipeng Ling, Shuliang Liu, Seonil Son +4Reasoning Trace Analysis Logical Reasoning
The Hong Kong University of Science and Technology (Guangzhou) · University of Alberta · Alberta Machine Intelligence Institute +3
Generating Effective CoT Traces for Mitigating Causal Hallucination
Apr 14, 2026Yiheng Zhao, Jun YanLLM Hallucination Mitigation Hallucination in Language Models
Concordia University, Montreal, Canada
Critical-CoT: A Robust Defense Framework against Reasoning-Level Backdoor Attacks in Large Language Models
Apr 12, 2026Vu Tuan Truong, Long Bao LeBackdoor Defense in LLMs LLM Backdoor Attacks
INRS, University of Quebec
Thinking Fast, Thinking Wrong: Intuitiveness Modulates LLM Counterfactual Reasoning in Policy Evaluation
Apr 12, 2026Yanjie HeCounterfactual Reasoning in Language Models Causal Reasoning in Language Models
Independent Researcher
Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty
Apr 11, 2026Chao Xue, Yao Wang, Mengqiao Liu +11Reward Modeling Efficient Language Model Reasoning
University of New South Wales, Australia · Tencent Hunyuan, China · Tencent Yuanbao, China +2
AI Achieves a Perfect LSAT Score
Apr 11, 2026Bonmu KuLLM Reasoning LLM Evaluation
How does Chain of Thought decompose complex tasks?
Apr 10, 2026Amrut Nadgir, Vijay Balasubramanian, Pratik ChaudhariCoT Reasoning Structured CoT Reasoning
University of Pennsylvania.
SeLaR: Selective Latent Reasoning in Large Language Models
Apr 9, 2026Renyu Fu, Guibo LuoCoT Reasoning Continuous Latent Reasoning
Guangdong Provincial Key Laboratory of Ultra High Definition Immersive Media Technology Shenzhen Graduate School, Peking University
Cognitive Loop of Thought: Reversible Hierarchical Markov Chain for Efficient Mathematical Reasoning
Apr 8, 2026Jia-Chen Zhang, Yu-Jie Xiong, Zheng ZhouStructured CoT Reasoning LLM Mathematical Reasoning
School of Computer Science and Technology, East China Normal University 3663 North Zhongshan Road, Shanghai, China · School of Electronic and Electrical Engineering, Shanghai University of Engineering Science 333 Longteng Road, Shanghai, China
Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability
Apr 8, 2026Qihan Ren, Peng Wang, Ruikun Cai +8LLM Fine-Tuning Cross-Domain Generalization
Shanghai Artificial Intelligence Laboratory · Shanghai Jiao Tong University · University of Science and Technology of China
Learning to Edit Knowledge via Instruction-based Chain-of-Thought Prompting
Apr 7, 2026Jinhu Fu, Yan Bai, Longzhu He +4Knowledge Editing Language Model Editing
Beijing University of Posts and Telecommunications · Peking University · Huawei Technologies Ltd. +1
ETR: Entropy Trend Reward for Efficient Chain-of-Thought Reasoning
Apr 7, 2026Xuan Xiong, Huan Liu, Li Gu +4Efficient Language Model Reasoning CoT Reasoning
University of Toronto · McMaster University · Concordia University +1
Strengthening Human-Centric Chain-of-Thought Reasoning Integrity in LLMs via a Structured Prompt Framework
Apr 6, 2026Jiling Zhou, Aisvarya Adeseye, Seppo Virtanen +2Structured Prompting CoT Reasoning
Department of Computing, University of Turku, 20014 Turku, Finland
InCoder-32B-Thinking: Industrial Code World Model for Thinking
Apr 3, 2026Jian Yang, Wei Zhang, Jiajun Wu +22LLM Code Generation Code Generation
Beihang University · IQuest Research · Shanghai Jiao Tong University +5

Apr 24, 2026J/K move · Enter open · S save

Calibration Drift Under Reasoning: How Chain-of-Thought Budgets Induce Overconfidence in Large Language Models

Prakul Sunil Hiremath, Harshit R. Hiremath

Department of Computer Science and Engineering Visvesvaraya Technological University, Belagavi · Department of Computer Science and Business System SG Balekundri Institute of Technology, Belagavi

Confidence Estimation in Language Models CoT Reasoning Language Model Calibration LLM Reliability LLM Uncertainty Estimation LLM Reasoning

Open paper AskPDF arXiv

The ability of large language models (LLMs) to express calibrated uncertainty is important for safe deployment. Chain-of-thought (CoT) reasoning is widely used to improve accuracy and reliability, but its effect on calibration is not fully understood. We show that this picture is incomplete: in some settings, increasing the reasoning budget beyond a task-specific threshold can cause models to become systematically overconfident, assigning high confidence to incorrect answers. We call this phenomenon Calibration Drift Under Reasoning (CDUR) and study it both theoretically and empirically. We define reasoning budget B and analyze conditions under which Expected Calibration Error ECE(B) follows a non-monotonic pattern: it first decreases as reasoning corrects errors, then increases as longer reasoning produces internally consistent but incorrect explanations. We propose a Hypothesis Lock-In model based on autoregressive generation to explain this behavior. We evaluate Llama-3.1-8B and Llama-3.3-70B on 47 reasoning-trap questions across four reasoning budgets and three seeds (1,368 API calls; 574 valid responses). The 8B model shows non-monotonic calibration behavior, while results for the 70B model are limited to baseline evaluation and are inconclusive for budget-dependent effects. We introduce CABStop, a calibration-aware stopping rule that halts reasoning when confidence diverges from an auxiliary accuracy estimate. These results suggest that increasing reasoning depth does not always improve reliability and should be monitored carefully.