SWE-Prime: Fewer Trajectories, Better Performance
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…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.
