Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Junjie, Ma, Guozheng, Liu, Shunyu, Hu, Zetian, Jing, Yongcheng, Lin, Ting-En, Li, Yongbin, Tao, Dacheng
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.18851
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909054949916672
author Zhang, Junjie
Ma, Guozheng
Liu, Shunyu
Hu, Zetian
Jing, Yongcheng
Lin, Ting-En
Li, Yongbin
Tao, Dacheng
author_facet Zhang, Junjie
Ma, Guozheng
Liu, Shunyu
Hu, Zetian
Jing, Yongcheng
Lin, Ting-En
Li, Yongbin
Tao, Dacheng
contents Recent advances in Reinforcement Learning (RL) have underscored its potential for incentivizing reasoning capabilities of Large Language Models (LLMs). However, existing step-level efforts suffer from costly annotations that limit domain coverage, while scalar scores further impose an information bottleneck, offering insufficient semantic bandwidth to improve intermediate decisions. Alternative language-critique approaches, which rely on frozen or external critics, provide richer textual feedback but lack the scalability needed for sustained policy improvement. In this work, we propose language-driven stepwise trajectory redirection, termed as STRIDE, a novel training framework that shifts process supervision from scalar rewards to learnable stepwise language feedback. Specifically, we co-train a generator and a generative verifier using only outcome-based rewards, eliminating external annotations, while delivering sustained policy improvement through jointly aligned verifier training. The verifier's stepwise language critiques explicitly localize and explain failures, enabling the generator to redirect reasoning trajectories at intermediate steps toward alternative decisions. The trajectory redirection design guarantees harmless policy improvement, even under noisy or suboptimal verifier feedback. Experiments on diverse reasoning benchmarks show that STRIDE significantly outperforms state-of-the-art baselines, as well as achieving breakthroughs on zero-pass-rate problems where scalar methods yield no learning signal in our ablation studies, demonstrating the effectiveness of learnable stepwise language feedback for enhancing LLM reasoning.
format Preprint
id arxiv_https___arxiv_org_abs_2605_18851
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning
Zhang, Junjie
Ma, Guozheng
Liu, Shunyu
Hu, Zetian
Jing, Yongcheng
Lin, Ting-En
Li, Yongbin
Tao, Dacheng
Machine Learning
Recent advances in Reinforcement Learning (RL) have underscored its potential for incentivizing reasoning capabilities of Large Language Models (LLMs). However, existing step-level efforts suffer from costly annotations that limit domain coverage, while scalar scores further impose an information bottleneck, offering insufficient semantic bandwidth to improve intermediate decisions. Alternative language-critique approaches, which rely on frozen or external critics, provide richer textual feedback but lack the scalability needed for sustained policy improvement. In this work, we propose language-driven stepwise trajectory redirection, termed as STRIDE, a novel training framework that shifts process supervision from scalar rewards to learnable stepwise language feedback. Specifically, we co-train a generator and a generative verifier using only outcome-based rewards, eliminating external annotations, while delivering sustained policy improvement through jointly aligned verifier training. The verifier's stepwise language critiques explicitly localize and explain failures, enabling the generator to redirect reasoning trajectories at intermediate steps toward alternative decisions. The trajectory redirection design guarantees harmless policy improvement, even under noisy or suboptimal verifier feedback. Experiments on diverse reasoning benchmarks show that STRIDE significantly outperforms state-of-the-art baselines, as well as achieving breakthroughs on zero-pass-rate problems where scalar methods yield no learning signal in our ablation studies, demonstrating the effectiveness of learnable stepwise language feedback for enhancing LLM reasoning.
title STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning
topic Machine Learning
url https://arxiv.org/abs/2605.18851