Reasoning Diversity

Follow this topic Ask about this topic

Momentum

1 paper in the last four weeks, against 2 the four weeks before. 0.0% of all new papers.

Jul 13Week of Sep 28

Latest papers 13

All topics

Selecting Diverse SFT Traces Improves Post-RL Generalization
Sep 27, 2026Dylan Zhang, Mingyuan Wu, Jinning LiReasoning Diversity RL for Language Model Reasoning
University of Illinois Urbana-Champaign, work done at Google · Google
Escaping Reasoning Basin Collapse with History-Biased Search
Sep 1, 2026Lu ChengInference-Time Search Reasoning Diversity
Department of Computer Science University of Illinois Chicago Chicago, IL 60607
Creative Generation via Multi-Agent Debate: Does Debate Suppress Diversity?
Sep 1, 2026Tien Anh Nguyen, Khanh-Binh Nguyen, Van Dai Do +2Reasoning Diversity Multi-Agent Debate
Deakin Applied Artificial Intelligence Initiative, Deakin University, Australia
BODHI: Do LLMs Branch Out and Discover Heterogeneous Inferences?
Aug 3, 2026Soumadeep Saha, Krish Sharma, Akshay Chaturvedi +1Reasoning Diversity RL for Language Model Reasoning
ANITI, Université de Toulouse · LINAGORA Labs
ReCo: Reweighting GRPO Against Distributional Concentration
Jul 29, 2026Junoh Park, Junseo Hwang, Wonguk Cho +1Reasoning Diversity RL for Language Model Reasoning
Graduate School of Data Science, Seoul National University
Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning
Jun 29, 2026Sangmook Lee, Minbeom Kim, Jeonghye Kim +3LLM Evaluation Reasoning Diversity
Seoul National University · KAIST
Understanding Diversity Collapse in RLVR via the Lens of Overtraining
Jun 13, 2026Suqin Yuan, Jinkun Chen, Jiyang Zheng +6Reinforcement Learning Reasoning Diversity
Sydney AI Centre, The University of Sydney · Southeast University · Microsoft +3
TruthSplit: Operationalizing Conditional Validity in Arguments Through Multi-Perspective Reasoning
Jun 8, 2026Benjamin Stieger, Maximilian Terberger, Thomas Huber +1Reasoning Diversity Computational Argumentation
University of St. Gallen
Diverse Thinking Schemata Elicit Better Reasoning in Large Language Models
Jun 8, 2026Xinyue Liang, Yizhe Yang, Yu Bai +3Reasoning Diversity RL for Language Model Reasoning
School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China
Hint-Guided Diversified Policy Optimization for LLM Reasoning
Jun 2, 2026Zhiyu Cao, Kaixin Wu, Mingjie Zhong +4Reasoning Diversity RL for Language Model Reasoning
School of Computer Science and Technology, Soochow University, Suzhou, China · Ant Group, Hangzhou, China
Why Do Reasoning Models Lose Coverage? The Role of Data and Forks in the Road
May 16, 2026Ngoc-Hieu Nguyen, Parshin Shojaee, Phuc Minh Nguyen +4LLM Fine-Tuning Reasoning Diversity
The Pennsylvania State University · Virginia Tech · VinUniversity
Diversity of Extensions in Abstract Argumentation
May 13, 2026Johannes K. Fichte, Markus Hecher, Yasir Mahmood +1Reasoning Diversity Computational Argumentation
Department of Computer and Information Science (IDA), Linköping University, Sweden · University of Potsdam, Germany · University of Artois, CNRS, UMR8188 (CRIL), France +1
Beyond Accuracy: Evaluating Strategy Diversity in LLM Mathematical Reasoning
May 10, 2026Xia Yang, Xuanyi Zhang, Hao Hu +1Mathematical Reasoning Benchmarks LLM Evaluation
University of Toronto · Upper Canada College · East China Normal University

Sep 27, 2026, cs.LGJ/K move · Enter open · S save

Selecting Diverse SFT Traces Improves Post-RL Generalization

Dylan Zhang, Mingyuan Wu, Jinning Li

University of Illinois Urbana-Champaign, work done at Google · Google

Reasoning Diversity RL for Language Model Reasoning Training Data Selection RL Post-Training LLM Reasoning

Open paper AskPDF arXiv

Verified solutions are not equally useful for preparing reasoning models for reinforcement learning (RL). We present a comprehensive study of route diversity, the variation in the sequences of reasoning steps in supervised fine-tuning (SFT) data, and propose a lightweight, rule-based fingerprint to select for it. From one pool at one budget, with matched training recipes and checkpoints, selecting diverse rather than similar routes improves post-RL problem coverage across puzzles and mathematics, including on problems harder than those seen in either training stage. In synthetic experiments, route-diverse SFT improves OLMo3-7B's pass@8 by 16.9 points on environments held out from SFT. In a single-model condition, where one model writes every candidate, diverse selection gains up to 6.2 points of mean pass@8 across 10 mathematics benchmarks. Pre-RL diagnostics suggest why: diverse SFT can produce both successful and failed attempts on more prompts despite slightly lower mean accuracy, giving group-relative RL more prompts with a learning signal. On 3 open-source corpora, our CPU-only selector, without model calls, outperforms more expensive alternatives in every comparison of mean post-RL performance. These results identify reasoning-route diversity as a practical criterion for selecting SFT data that better prepares models for RL.

Momentum

Related topics

Latest papers 13