Reproduce “Exploring Large Language Models in Resolving Environment-Related Crash Bugs: Localizing and Repairing” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xueying Du, Mingwei Liu, Hanlin Wang, Juntao Li, Xin Peng, Yiling Lou
IntDiagSolver uses targeted questions to gather missing environmental details and refine crash repairs through dialogue. On 30 selected environment-related cases, it repaired 16, compared with 0 for a basic GPT-3.5 prompt, and localized 25, compared with 12. Cases counted only when two of three runs were correct; proposed fixes were not independently executed and verified.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.