Raw SWE-agent Trajectories and Predictions for LivePlan Evaluation on SWE-bench Verified and SWE-bench Pro
收藏官方服务：
资源简介：
This dataset contains the raw SWE-agent execution trajectories and prediction outputs used to evaluate LivePlan on SWE-bench Verified and SWE-bench Pro. It includes runs under LivePlan’s online monitoring and advising interventions, as well as the corresponding baseline configurations. The data supports analysis and reproduction of the execution behaviors and task-resolution results reported in the LivePlan study.
提供机构：
Zenodo创建时间：
2026-07-21
