AI Paper Atlas

Research map / Large language models / Agents, code and math

Pull Requests as a Training Signal for Repo-Level Code Editing

Qinglin Zhu, Tianyu Chen, Shuai Lu, Lei Ji, Runcong Zhao, Murong Ma, Xiangxiang Dai, Yulan He, Lin Gui, Peng CHENG, Yeyun Gong

ICML 2026 · Large language models › Agents, code and math · approach: Software engineering agents · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add repo-level mid-training (StarCoder2-style) before SFT (ablation)Qwen2.5-Coder-32B · SWE-bench LitePass@1: 11.3 -> 15.7Improved
Replace StarCoder2-style data with Clean-PR-train (ablation)Qwen2.5-Coder-32B · SWE-bench VerifiedPass@1: 20.4 -> 30.6Improved
Use Search/Replace instead of Unified Diff (ablation)Clean-PR (Python) · SWE-bench VerifiedPass@1: 24.4 -> 27.8Improved
Add linked issue descriptions to PR context (ablation)Clean-PR (Python) · SWE-bench VerifiedPass@1: 25.7 -> 27.8Improved
Train on all 12 languages instead of Python only (ablation)Clean-PR · SWE-bench VerifiedPass@1: 27.8 -> 30.6Improved
Add error-driven augmentation to SFT (ablation)Clean-PR (All languages) · SWE-bench LitePass@1: 21.8 -> 24.3Improved
Add error-driven augmentation to SFT (ablation)Clean-PR (All languages) · SWE-bench VerifiedPass@1: 27.4 -> 30.6Improved
Add error-driven augmentation to SFT (ablation)Clean-PR (Python) · SWE-bench LiteFile accuracy: 86.7 -> 86.3No clear effect
Mid-train on StarCoder2-style diffs (ablation)Qwen2.5-Coder-32B · HumanEvalPass@1: 54.1% -> 47.6%Hurt
Mid-train on Clean-PR Search/Replace data (ablation)Qwen2.5-Coder-32B · HumanEvalPass@1: reaching 59.8% (+5.7%)Improved
Mid-train on Clean-PR Search/Replace data (ablation)Qwen2.5-Coder-32B · LiveCodeBenchPass@1: 29.0% -> 32.6%Improved
Sample k=10 patches instead of k=1Clean-PR (All languages) · SWE-bench VerifiedPass@k: 30.6 -> 41.5Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy