TreeRPO: Tree Relative Policy Optimization

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Yang, Zhicheng, Guo, Zhijiang, Huang, Yinya, Liang, Xiaodan, Wang, Yiwei, Tang, Jing
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908562544918528
author Yang, Zhicheng
Guo, Zhijiang
Huang, Yinya
Liang, Xiaodan
Wang, Yiwei
Tang, Jing
author_facet Yang, Zhicheng
Guo, Zhijiang
Huang, Yinya
Liang, Xiaodan
Wang, Yiwei
Tang, Jing
contents Large Language Models (LLMs) have shown remarkable reasoning capabilities through Reinforcement Learning with Verifiable Rewards (RLVR) methods. However, a key limitation of existing approaches is that rewards defined at the full trajectory level provide insufficient guidance for optimizing the intermediate steps of a reasoning process. To address this, we introduce \textbf{\name}, a novel method that estimates the mathematical expectations of rewards at various reasoning steps using tree sampling. Unlike prior methods that rely on a separate step reward model, \name directly estimates these rewards through this sampling process. Building on the group-relative reward training mechanism of GRPO, \name innovatively computes rewards based on step-level groups generated during tree sampling. This advancement allows \name to produce fine-grained and dense reward signals, significantly enhancing the learning process and overall performance of LLMs. Experimental results demonstrate that our \name algorithm substantially improves the average Pass@1 accuracy of Qwen-2.5-Math on test benchmarks, increasing it from 19.0\% to 35.5\%. Furthermore, \name significantly outperforms GRPO by 2.9\% in performance while simultaneously reducing the average response length by 18.1\%, showcasing its effectiveness and efficiency. Our code will be available at \href{https://github.com/yangzhch6/TreeRPO}{https://github.com/yangzhch6/TreeRPO}.
format Preprint
id arxiv_https___arxiv_org_abs_2506_05183
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TreeRPO: Tree Relative Policy Optimization
Yang, Zhicheng
Guo, Zhijiang
Huang, Yinya
Liang, Xiaodan
Wang, Yiwei
Tang, Jing
Machine Learning
Artificial Intelligence
Large Language Models (LLMs) have shown remarkable reasoning capabilities through Reinforcement Learning with Verifiable Rewards (RLVR) methods. However, a key limitation of existing approaches is that rewards defined at the full trajectory level provide insufficient guidance for optimizing the intermediate steps of a reasoning process. To address this, we introduce \textbf{\name}, a novel method that estimates the mathematical expectations of rewards at various reasoning steps using tree sampling. Unlike prior methods that rely on a separate step reward model, \name directly estimates these rewards through this sampling process. Building on the group-relative reward training mechanism of GRPO, \name innovatively computes rewards based on step-level groups generated during tree sampling. This advancement allows \name to produce fine-grained and dense reward signals, significantly enhancing the learning process and overall performance of LLMs. Experimental results demonstrate that our \name algorithm substantially improves the average Pass@1 accuracy of Qwen-2.5-Math on test benchmarks, increasing it from 19.0\% to 35.5\%. Furthermore, \name significantly outperforms GRPO by 2.9\% in performance while simultaneously reducing the average response length by 18.1\%, showcasing its effectiveness and efficiency. Our code will be available at \href{https://github.com/yangzhch6/TreeRPO}{https://github.com/yangzhch6/TreeRPO}.
title TreeRPO: Tree Relative Policy Optimization
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.05183