Boosting Open-Source LLMs for Program Repair via Reasoning Transfer and LLM-Guided Reinforcement Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Tang, Xunzhu, Klein, Jacques, Bissyandé, Tegawendé F.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908393179971584
author Tang, Xunzhu
Klein, Jacques
Bissyandé, Tegawendé F.
author_facet Tang, Xunzhu
Klein, Jacques
Bissyandé, Tegawendé F.
contents Several closed-source LLMs have consistently outperformed open-source alternatives in program repair tasks, primarily due to their superior reasoning capabilities and extensive pre-training. This paper introduces Repairity, a novel three-stage methodology that significantly narrows this performance gap through reasoning extraction and reinforcement learning. Our approach: (1) systematically filters high-quality reasoning traces from closed-source models using correctness verification, (2) transfers this reasoning knowledge to open-source models via supervised fine-tuning, and (3) develops reinforcement learning with LLM-based feedback to further optimize performance. Empirical evaluation across multiple program repair benchmarks demonstrates that Repairity improves the performance of Qwen2.5-Coder-32B-Instruct, a base open source LLM, by 8.68\% on average, reducing the capability gap with Claude-Sonnet3.7, a state-of-the-art closed-source model, from 10.05% to 1.35%. Ablation studies confirm that both reasoning extraction and LLM-guided reinforcement learning contribute significantly to these improvements. Our methodology generalizes effectively to additional code-related tasks, enabling organizations to leverage high-quality program repair capabilities while maintaining the customizability, transparency, and deployment flexibility inherent to open-source models.
format Preprint
id arxiv_https___arxiv_org_abs_2506_03921
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Boosting Open-Source LLMs for Program Repair via Reasoning Transfer and LLM-Guided Reinforcement Learning
Tang, Xunzhu
Klein, Jacques
Bissyandé, Tegawendé F.
Software Engineering
Several closed-source LLMs have consistently outperformed open-source alternatives in program repair tasks, primarily due to their superior reasoning capabilities and extensive pre-training. This paper introduces Repairity, a novel three-stage methodology that significantly narrows this performance gap through reasoning extraction and reinforcement learning. Our approach: (1) systematically filters high-quality reasoning traces from closed-source models using correctness verification, (2) transfers this reasoning knowledge to open-source models via supervised fine-tuning, and (3) develops reinforcement learning with LLM-based feedback to further optimize performance. Empirical evaluation across multiple program repair benchmarks demonstrates that Repairity improves the performance of Qwen2.5-Coder-32B-Instruct, a base open source LLM, by 8.68\% on average, reducing the capability gap with Claude-Sonnet3.7, a state-of-the-art closed-source model, from 10.05% to 1.35%. Ablation studies confirm that both reasoning extraction and LLM-guided reinforcement learning contribute significantly to these improvements. Our methodology generalizes effectively to additional code-related tasks, enabling organizations to leverage high-quality program repair capabilities while maintaining the customizability, transparency, and deployment flexibility inherent to open-source models.
title Boosting Open-Source LLMs for Program Repair via Reasoning Transfer and LLM-Guided Reinforcement Learning
topic Software Engineering
url https://arxiv.org/abs/2506.03921