Reproduce “An Empirical Study on Improving LLM-Based Issue Classification through Domain-Informed Analysis” 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
Anubhav Gupta, Tanner Wright, Gema Rodríguez-Pérez
The authors evaluate a multi-stage method for classifying closed GitHub issues as Intrinsic Bug, Extrinsic Bug, or Not a Bug. It first assesses whether a genuine defect exists, then identifies its origin, with recovery and domain-informed guidance. On 428 labeled NPM issues, macro F1 and Extrinsic Bug recall improve for five of six models; external-cause reports remain difficult.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.