Research

Paper

TESTING March 09, 2026

SWE-Fuse: Empowering Software Agents via Issue-free Trajectory Learning and Entropy-aware RLVR Training

Authors

Xin-Cheng Wen, Binbin Chen, Haoxuan Lan, Hang Yu, Peng Di, Cuiyun Gao

Abstract

Large language models (LLMs) have transformed the software engineering landscape. Recently, numerous LLM-based agents have been developed to address real-world software issue fixing tasks. Despite their state-of-the-art performance, Despite achieving state-of-the-art performance, these agents face a significant challenge: \textbf{Insufficient high-quality issue descriptions.} Real-world datasets often exhibit misalignments between issue descriptions and their corresponding solutions, introducing noise and ambiguity that mislead automated agents and limit their problem-solving effectiveness. We propose \textbf{\textit{SWE-Fuse}}, an issue-description-aware training framework that fuses issue-description-guided and issue-free samples for training SWE agents. It consists of two key modules: (1) An issue-free-driven trajectory learning module for mitigating potentially misleading issue descriptions while enabling the model to learn step-by-step debugging processes; and (2) An entropy-aware RLVR training module, which adaptively adjusts training dynamics through entropy-driven clipping. It applies relaxed clipping under high entropy to encourage exploration, and stricter clipping under low entropy to ensure training stability. We evaluate SWE-Fuse on the widely studied SWE-bench Verified benchmark shows to demonstrate its effectiveness in solving real-world software problems. Specifically, SWE-Fuse outperforms the best 8B and 32B baselines by 43.0\% and 60.2\% in solve rate, respectively. Furthermore, integrating SWE-Fuse with test-time scaling (TTS) enables further performance improvements, achieving solve rates of 49.8\% and 65.2\% under TTS@8 for the 8B and 32B models, respectively.

Metadata

arXiv ID: 2603.07927
Provider: ARXIV
Primary Category: cs.SE
Published: 2026-03-09
Fetched: 2026-03-10 05:43

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.07927v1</id>\n    <title>SWE-Fuse: Empowering Software Agents via Issue-free Trajectory Learning and Entropy-aware RLVR Training</title>\n    <updated>2026-03-09T03:47:10Z</updated>\n    <link href='https://arxiv.org/abs/2603.07927v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.07927v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Large language models (LLMs) have transformed the software engineering landscape. Recently, numerous LLM-based agents have been developed to address real-world software issue fixing tasks. Despite their state-of-the-art performance, Despite achieving state-of-the-art performance, these agents face a significant challenge: \\textbf{Insufficient high-quality issue descriptions.} Real-world datasets often exhibit misalignments between issue descriptions and their corresponding solutions, introducing noise and ambiguity that mislead automated agents and limit their problem-solving effectiveness. We propose \\textbf{\\textit{SWE-Fuse}}, an issue-description-aware training framework that fuses issue-description-guided and issue-free samples for training SWE agents. It consists of two key modules: (1) An issue-free-driven trajectory learning module for mitigating potentially misleading issue descriptions while enabling the model to learn step-by-step debugging processes; and (2) An entropy-aware RLVR training module, which adaptively adjusts training dynamics through entropy-driven clipping. It applies relaxed clipping under high entropy to encourage exploration, and stricter clipping under low entropy to ensure training stability. We evaluate SWE-Fuse on the widely studied SWE-bench Verified benchmark shows to demonstrate its effectiveness in solving real-world software problems. Specifically, SWE-Fuse outperforms the best 8B and 32B baselines by 43.0\\% and 60.2\\% in solve rate, respectively. Furthermore, integrating SWE-Fuse with test-time scaling (TTS) enables further performance improvements, achieving solve rates of 49.8\\% and 65.2\\% under TTS@8 for the 8B and 32B models, respectively.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.SE'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-09T03:47:10Z</published>\n    <arxiv:comment>19 pages</arxiv:comment>\n    <arxiv:primary_category term='cs.SE'/>\n    <author>\n      <name>Xin-Cheng Wen</name>\n    </author>\n    <author>\n      <name>Binbin Chen</name>\n    </author>\n    <author>\n      <name>Haoxuan Lan</name>\n    </author>\n    <author>\n      <name>Hang Yu</name>\n    </author>\n    <author>\n      <name>Peng Di</name>\n    </author>\n    <author>\n      <name>Cuiyun Gao</name>\n    </author>\n  </entry>"
}