QiMeng-PRepair: Precise Code Repair via Edit-Aware Reward Optimization
This addresses the issue of imprecise code repairs for developers using LLMs, though it is an incremental improvement on existing repair methods.
The paper tackles the problem of over-editing in large language models for code repair, where excessive modifications overwrite correct code, by introducing a precise repair task that maximizes reuse of correct code. The proposed PRepair framework improves repair precision by up to 31.4% under a joint correctness and extent metric and increases decoding throughput with speculative editing.
Large Language Models (LLMs) achieve strong program repair performance but often suffer from over-editing, where excessive modifications overwrite correct code and hinder bug localization. We systematically quantify its impact and introduce precise repair task, which maximizes reuse of correct code while fixing only buggy parts. Building on this insight, we propose PRepair, a framework that mitigates over-editing and improves repair accuracy. PRepair has two components: Self-Breaking, which generates diverse buggy programs via controlled bug injection and min-max sampling, and Self-Repairing, which trains models with Edit-Aware Group Relative Policy Optimization (EA-GRPO) using an edit-aware reward to encourage minimal yet correct edits. Experiments show that PRepair improves repair precision by up to 31.4% under $\mathrm{fix}_1@1$, a metric that jointly considers repair correctness and extent, and significantly increases decoding throughput when combined with speculative editing, demonstrating its potential for precise and practical code repair.