SWE-Lego: Pushing the Limits of Supervised Fine-tuning for Software Issue Resolving
Paper • 2601.01426 • Published • 25 •
4
🤝 Open to Collab
Yuxin Jiang
YuxinJiang
AI & ML interests
NLP
Recent Activity
updated a model 10 days ago
SWE-Lego/qwen3_8b_glm5_swerebench_oraclesolved_oh_sdk_512_gbs64pbs1acc8_lr1e-4_epo3 published a model 10 days ago
SWE-Lego/qwen3_8b_glm5_swerebench_oraclesolved_oh_sdk_512_gbs64pbs1acc8_lr1e-4_epo3 upvoted a paper 17 days ago
What Makes Interaction Trajectories Effective for Training Terminal Agents?Organizations
about openswe_traj.jsonl
#3 opened 3 months ago
by YuxinJiang
Add task category and size categories metadata
#2 opened 6 months ago
by 
nielsr
Add task category and improve metadata
#1 opened 6 months ago
by nielsr
commented a
paper 6 months ago