Software Engineering Empirical Research Radar

SWE-bench

Dataset and benchmark summary for software-engineering research.

Category

LLM agents / real GitHub issue resolution

Local Status

papers-or-notes-found

What It Is

A benchmark of real GitHub issues where systems must modify repositories to resolve software problems.

Local Evidence

Evidence TypeCount
Local PDF filename matches2
Local abstract filename matches0
Local fulltext filename matches2
Local notes / metadata mentions2
Paper-title mentions in snapshot2

Local Dataset Directories