Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities under the widely adopted SFT+RLVR paradigm, which first performs Supervised Fine-Tuning (SFT) on human-annotated reasoning trajectories (rationales) to establish initial reasoning behaviors, then applies Reinforcement Learning with Verifiable Rewards (RLVR) to optimize the model using verifiable signals without golden rationales. However, annotating high-quality rationales for the SFT stage remains prohibitively expensive. This paper investigates when and how rationale annotation costs can be substantially reduced without compromising reasoning performance. We identify a broad class of problems, termed patterned reasoning tasks, where reasoning follows a fixed, procedural strategy consistent across instances. Although instances vary in content such as domain knowledge, factual information, or numeric values, the solution derives from applying a shared reasoning pattern. We argue that the success of SFT+RLVR on such tasks primarily stems from its ability to enable models to internalize these reasoning patterns. Using numerical semantic matching as a representative task, we provide both causal and behavioral evidence showing that reasoning patterns rather than the quantity or quality of rationales are the key determinant of performance. Building on these insights, we propose Pattern-Aware LLMs as Rationale AnnOtators (PARO), a simple yet effective framework that enables LLMs to generate rationales aligned with task-specific reasoning patterns without requiring human rationale annotations. Experiments show that PARO-generated rationales achieve comparable SFT+RLVR performance to human rationales that are 10 times larger. These results suggest that large-scale human rationale annotations can be replaced with LLM-based automatic annotations requiring only limited human supervision over reasoning patterns.
Verbosity-Aware Rationale Reduction: Effective Reduction of Redundant Rationale via Principled Criteria
POSTECH;Seoul National University(首尔国立大学);University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Kyungpook National University(庆北国立大学);LG AI Research(LG AI研究院)
Knowledge-Augmented Multimodal Clinical Rationale Generation for Disease Diagnosis with Small Language Models
Hong Kong Baptist University(香港 Baptist 大学);Shanxi University(山西大学);Shanghai Institute for Advanced Study of Zhejiang University(浙江大学上海先进研究院);Hong Kong University of Science and Technology(香港科技大学);The University of Manchester(曼彻斯特大学)
Diverge to Induce Prompting: Multi-Rationale Induction for Zero-Shot Reasoning
Department of Computer Science and Information Engineering, National Taiwan University, Taiwan(国家台湾大学计算机科学与信息工程系);Institute of Information Science, Academia Sinica, Taiwan(学术院信息科学研究所);AI Research Center (AINTU), National Taiwan University, Taiwan(国家台湾大学人工智能研究中心(AINTU))
Rationale-Grounded In-Context Learning for Time Series Reasoning with Multimodal Large Language Models
The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州));China Mobile (Jiangxi) Virtual Reality Technology Co., Ltd.(中国移动(江西)虚拟现实技术有限公司);School of Computer and Information Technology, Beijing Jiaotong University(北京交通大学计算机与信息学院);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所)
Audit Before You Distill: A Controlled Workflow for Reasoning Transfer into Compact Medical Language Models
Einstein Global Advanced Technologies for Equity (E-GATE), Hospital Israelita Albert Einstein(以色列阿尔伯特·爱因斯坦医院 爱因斯坦全球公平先进技术中心(E-GATE));Institute of Science and Technology Itaú (ICTi), Itaú Unibanco(伊塔乌联合银行 伊塔乌科学技术研究所(ICTi));School of Sciences, São Paulo State University(圣保罗州立大学 理学院)
MIND: Multi-rationale INtegrated Discriminative Reasoning Framework for Multi-modal Large Models
Shenyang Institute of Automation, Chinese Academy of Sciences(中国科学院沈阳自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);Peking University(北京大学);MMLab, CUHK(CUHK多模态实验室)
From Prediction to Justification: Aligning Sentiment Reasoning with Human Rationale via Reinforcement Learning
School of Computer Science and Technology, East China Normal University(东华大学计算机科学与技术学院);Shanghai Qiji Zhifeng Co., Ltd.(上海启智丰科技有限公司);Ocean University of China(中国海洋大学)
Learning to Rationalize for Nonmonotonic Reasoning with Distant Supervision
University of California, Santa Cruz(加州大学圣克鲁兹分校);Allen Institute for AI(艾伦人工智能研究所);Paul G. Allen School of Computer Science & Engineering, University of Washington(华盛顿大学保罗·G·艾伦计算机科学与工程学院);University of Maryland, College Park(马里兰大学学院公园分校)
Boosting the Power of Small Multimodal Reasoning Models to Match Larger Models with Self-Consistency Training
Zhejiang University(浙江大学);Westlake University(西湖大学);Shenyang Institute of Computing Technology, Chinese Academy of Sciences(中国科学院沈阳自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);Institute of Advanced Technology, Westlake Institute for Advanced Study(西湖高等研究院先进技术研究院)
Image-of-Thought Prompting for Visual Reasoning Refinement in Multimodal Large Language Models
School of Engineering, Westlake University(西湖大学工程学院);City University of Hong Kong(香港城市大学);Research Center for Industries of the Future, Westlake University(西湖大学未来产业研究中心);Johns Hopkins University(约翰斯·霍普金斯大学)
GNN2R: Weakly-Supervised Rationale-Providing Question Answering over Knowledge Graphs
University of Zurich(苏黎世大学);SIB Swiss Institute of Bioinformatics(瑞士生物信息学研究所);School of Computing, Dublin City University(都柏林城市大学计算机学院);Ellis Institute Finland(芬兰埃利斯研究所);Åbo Akademi University(阿博阿卡迪米大学)
College of William and Mary(威廉与玛丽学院);University of Toronto(多伦多大学);University Health Network(大学健康网络);KU Leuven(鲁汶大学);Bucknell University(巴克内尔大学);University of Cincinnati(辛辛那提大学);University of Alabama at Birmingham(伯明翰大学)
The Impact of Reasoning Step Length on Large Language Models
Rutgers University(罗格斯大学);University of Liverpool(利物浦大学);Northwestern University(西北大学);New Jersey Institute of Technology(新泽西理工学院);University of Exeter(埃克塞特大学)