Adaptive Reasoning in Language Models

Latest papers

49 papers22 papers in the last four weeks, against 2 the four weeks before. 0.1% of all new papers.

Calibration-Free Early Stopping for Chain-of-Thought Reasoning via E-processes
Oct 18, 2026Dong-Jae Lee, Gyojin Han, Junmo KimEarly Stopping for Language Model Reasoning Adaptive Reasoning in Language Models
KAIST
SteadyThought: Mitigating LLM Under-Thinking via Thought-Level Preference Optimization
Oct 18, 2026Zhenyue Peng, Lin Yang, Jie Wang +7Efficient Language Model Reasoning Adaptive Reasoning in Language Models
Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei 230031, China · University of Science and Technology of China, Hefei 230026, China · Shanghai Artificial Intelligence Laboratory, Shanghai, China +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 Models Efficient Language Model Reasoning
LG AI Research · UNIST · University of Illinois Chicago
When Should Agents Think? Adaptive Reasoning via Cross-Turn Estimation
Oct 8, 2026Yiruo Cheng, Shen Huang, Xiaoshuai Song +5Adaptive Reasoning in Language Models Agentic Reasoning
Gaoling School of Artificial Intelligence, Renmin University of China · Alibaba Token Hub, Alibaba Group
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 Models RL for Language Model Reasoning
Southern University of Science and Technology · The Hong Kong University of Science and Technology · Shanghai Artificial Intelligence Laboratory +1
Hunyuan-A13B Technical Report
Sep 23, 2026Tencent Hunyuan Team, Ao Liu, Botong Zhou +72Adaptive Reasoning in Language Models Efficient Language Model Reasoning
Selective State-Space Adaptation and Retrieval for Language Model Reasoning
Jul 21, 2026Atahan Dokme, Larry HeckRetrieval-Augmented Generation Selective SSMs
AI Virtual Assistant (AVA) Lab Georgia Institute of Technology
Switch-Reasoner: Learn When to Think in Multitask Mixtures via Reinforcement Learning
Jul 9, 2026Yiyang Fang, Pei Fu, Jinjie Li +7Large Reasoning Models Efficient Multimodal Inference
Wuhan University · The Hong Kong University of Science and Technology · Xiaomi Inc +2
SuCo: Sufficiency-guided Continuous Adaptive Reasoning
Jun 16, 2026Jiahao Wang, Bingyu Liang, Chenhao Hu +5Adaptive Reasoning in Language Models RL for Language Models
Harbin Institute of Technology (Shenzhen), Guangdong, China. · TeleAI of China Telecom, China.
Tyler: Typed Latent Reasoning for Language Models -- When to Think, What to Compute, and How Much to Allocate
Jun 15, 2026Hanyu Lin, Min Cai, Jiawei Wen +1Latent Reasoning LLM Reasoning
1Shenzhen University · University of Alberta
AdaptR1: Reinforcement Learning Based Adaptive Interleaved Thinking in Multi-hop Question Answering
May 29, 2026Yuxin Wang, Jiahao Lu, Qifeng Wu +5Multi-Hop QA Adaptive Reasoning in Language Models
1Computer Science, Fudan University · Institute of Modern Languages and Linguistics, Fudan University · 3Shanghai Innovation Institute +1
x1: Learning to Think Adaptively Across Languages and Cultures
Apr 18, 2026Yangfan Ye, Xiaocheng Feng, Xiachong Feng +8Multilingual Language Models Cross-Lingual Reasoning in Language Models
Harbin Institute of Technology · Peng Cheng Laboratory · The University of Hong Kong +1
HiRO-Nav: Hybrid ReasOning Enables Efficient Embodied Navigation
Apr 9, 2026He Zhao, Yijun Yang, Zichuan Lin +2Adaptive Reasoning in Language Models Embodied Navigation
Nanyang Technological University · Tencent · Independent Researcher
CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning
Mar 9, 2026Siye Wu, Jian Xie, Yikai Zhang +1Adaptive Reasoning in Language Models Adaptive Computation
♠Fudan University
Learning to Diagnose and Correct Moral Errors in Large Language Models
Jan 6, 2026Bocheng Chen, Xi Chen, Han Zi +6Moral Reasoning in Language Models LLM Self-Correction
University of Mississippi · Nanyang Technological University · Indiana University Indianapolis +4
REAP: Enhancing RAG with Recursive Evaluation and Adaptive Planning for Multi-Hop Question Answering
Nov 13, 2025Yijie Zhu, Haojie Zhou, Wanting Hong +2Multi-Hop QA Retrieval-Augmented Reasoning
The School of Artificial Intelligence and Computer Science, Jiangnan University
Enhancing Speech Large Language Models through Reinforced Behavior Alignment
Aug 25, 2025Yansong Liu, Yuan LiuSpeech Language Models Language Model Robustness
Xi’an Jiaotong-Liverpool University · The Chinese University of Hong Kong
Fast, Slow, and Tool-augmented Thinking for LLMs: A Review
Aug 17, 2025Xinda Jia, Jinpeng Li, Zezhong Wang +6Adaptive Reasoning in Language Models LLM Tool Use
School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China · Huawei Technologies Co., Ltd., Beijing 100084, China · Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Hong Kong 999077, China +3
Optimizing Anytime Reasoning via Budget Relative Policy Optimization
May 19, 2025Penghui Qi, Zichen Liu, Tianyu Pang +3Efficient Language Model Reasoning LLM Reasoning
Sea AI Lab · National University of Singapore

Oct 18, 2026J/K move · Enter open · S save

Calibration-Free Early Stopping for Chain-of-Thought Reasoning via E-processes

Dong-Jae Lee, Gyojin Han, Junmo Kim

KAIST

Early Stopping for Language Model Reasoning Adaptive Reasoning in Language Models Efficient Language Model Reasoning Sequential Hypothesis Testing

Open paper AskPDF arXiv

Reasoning large language models (LLMs) benefit from test-time scaling, motivating adaptive policies that spend tokens when they help and stop early otherwise. Existing rules halt when a confidence or stability signal crosses a threshold, but such thresholds rarely transfer across models and datasets, and validation-calibrated rules lose their guarantee under shifts in test-time horizon or signal dynamics. We reframe adaptive stopping as sequential evidence accumulation along the reasoning trajectory rather than threshold calibration on held-out data. Our framework runs two e-processes on a shared per-chunk signal: a commit stream tracks the signal's position relative to its slow trend, and an abstain stream tracks its signed velocity, making the two anti-correlated by construction. Under signal-level nulls on the per-chunk increments, Ville's inequality yields an anytime-valid signal-level false-alarm bound at user-specified significance levels, with no held-out calibration. Across various model--dataset combinations, the framework matches or exceeds calibration-based stopping on the token--accuracy Pareto front using zero calibration data, and the same significance level transfers unchanged across cells where calibration becomes infeasible.