Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Diverse Thinking Schemata Elicit Better Reasoning in Large Language Models

About

Large reasoning models (LRMs) have attracted increasing attention for their ability to solve complex mathematical problems by generating extended reasoning chains. In this work, we focus on two critical yet underexplored aspects of the reasoning process: reasoning transitions capturing the distinct transitions between reasoning steps and answer candidates reflecting the variety of solution paths produced by the model. We collectively define these two aspects as thinking schemata. We observe a correlation between the diversity of thinking schemata and model performance, which motivates us to enhance diversity as a means to further improve reasoning potential. To this end, we propose Diverse Schemata Policy Optimization (DiScO), a framework that first endows the model with schemata awareness, then encourages diversity through reinforcement learning, and further promotes diverse reasoning at inference time. Experiments on multiple mathematical reasoning benchmarks demonstrate that DiScO consistently outperforms standard group relative policy optimization. Beyond accuracy, human-annotated analyses show that DiScO substantially improves the model's ability to recover from erroneous initial attempts. Overall, our work suggests the important role that diversity of the thinking schemata plays and points to scaling along the diversity dimension as a promising research direction.

Xinyue Liang, Yizhe Yang, Yu Bai, Bin Xu, Jiawei Li, Yang Gao• 2026

Related benchmarks

TaskDatasetResultRank
ReasoningMMLU-Pro--
264
ReasoningGPQA
Pass@164
92
Mathematical ReasoningAIME 2025
Pass@1 Accuracy66.7
79
ReasoningARC Challenge
P@188
43
Competitive Mathematical ReasoningAIME 2024
Pass@1 Accuracy86.7
37
Competitive Mathematical ReasoningAMC 2023
Pass@1 Accuracy97.5
37
Mathematical ReasoningAMC 2023
Pass@8 Accuracy93.7
5
Showing 7 of 7 rows

Other info

Follow for update