paper-with-me

Papers

On the Direction of RLVR Updates for LLM Reasoning: Identification and Exploitation

2026-03-23 · Kexin Huang, Haoming Meng, Junkang Wu, Jinda Lu, Chiyu Ma, Ziqian Chen, Xue Wang, Bolin Ding, Jiancan Wu, Xiang Wang, Xiangnan He, Guoyin Wang, Jingren Zhou arxiv

Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning capabilities of large language models. While existing analyses identify that RLVR-induced changes are sparse, they primarily focus on the \textbf{magnitude} of these updates, largely overlooking their \textbf{direction}. In this work, we argue that the direction of updates is a more critical lens for understanding RLVR's effects, which can be captured by the signed, token-level log probability difference $Δ\log p$ between the base and final RLVR models. Through statistical analysis and token-replacement interventions, we demonstrate that $Δ\log p$ more effectively identifies sparse, yet reasoning-critical updates than magnitude-based metrics (\eg divergence or entropy). Building on this insight, we propose two practical applications: (1) a \textit{test-time extrapolation} method that amplifies the policy along the learned $Δ\log p$ direction to improve reasoning accuracy without further training; (2) a \textit{training-time reweighting} method that focuses learning on low-probability (corresponding to higher $Δ\log p$) tokens, which improves reasoning performance across models and benchmarks. Our work establishes the direction of change as a key principle for analyzing and improving RLVR.

📄 PDF Abstract BibTeX arXiv:2603.22117

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Controllable Exploration in Hybrid-Policy RLVR for Multi-Modal Reasoning

2026-02-22 · Zhuoxu Huang, Mengxi Jia, Hao Sun, Xuelong Li 외 arxiv

Reinforcement Learning with verifiable rewards (RLVR) has emerged as a primary learning paradigm for enhancing the reasoning capabilities of multi-modal large language models (MLLMs). However, during RL training, the eno…

Reinforcement Learning

Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning

2025-06-02 · Shenzhi Wang, Le Yu, Chang Gao, Chujie Zheng 외

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs), while its mechanisms are not yet well understood. In this …

Where Hindsight Credit Can Reside: A Signed-Capacity View of Token Updates in RLVR

2026-04-13 · Yuhang He, Haodong Wu, Siyi Liu, Hongyu Ge 외 arxiv

Reinforcement Learning with Verifiable Rewards (RLVR) improves the reasoning ability of Large Language Models (LLMs), but sparse outcome rewards make token-level credit assignment difficult. We study token-level credit a…

Reinforcement LearningMathematical Reasoning

Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models

2025-08-14 · Zhipeng Chen, Xiaobo Qin, Youbin Wu, Yue Ling 외 arxiv

Reinforcement learning with verifiable rewards (RLVR), which typically adopts Pass@1 as the reward, has faced the issues in balancing exploration and exploitation, causing policies to prefer conservative actions, converg…

Reinforcement Learning

Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward

2025-12-18 · Peter Chen, Xiaopeng Li, Ziniu Li, Wotao Yin 외 arxiv

This paper examines the exploration-exploitation trade-off in reinforcement learning with verifiable rewards (RLVR), a framework for improving the reasoning of Large Language Models (LLMs). Recent studies suggest that RL…

Reinforcement LearningMathematical Reasoning