SOURCE-LINKED INTELLIGENCE

SWE-Prime: Fewer Trajectories, Better Performance

arXiv · AI, language, vision and robotics · article · Aug 27, 2026 · UTC

To improve large language models' ability to resolve real-world software issues, prior work has focused on constructing large-scale agent trajectory datasets and performing supervised fine-tuning (SFT) on successful trajectories. However, task success does not guarantee high-quality supervision: successful trajectories may still contain ineffective, redundant, or risky steps. Directly using such trajectories for SFT can introduce noisy supervision and encourage models to imitate undesirable problem-solving behaviors. Therefore, we propose SWE-Prime, a multi-granularity, two-stage SFT data sele

Read original source ↗ Open in workspace

recordType

paper

region

Global