Reproduce “RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents” 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
Mingxuan Zhang, Xiaowen Wang, Anupma Sharan, Zhengyi Chen, Chenyu Diana Zhang, Shanshan Yang, Chittibabu Pacharu
RAFT indexes closed support cases as ordered timelines of symptoms, hypotheses, findings, and resolutions, then retrieves matching entries as a new case develops. It returns the parent case and identifies the matched entry. On a synthetic benchmark, initial-symptom-only queries yielded 84.2% Case Hit versus 67.3% for vanilla RAG. The evaluation tests retrieval, not diagnosis or resolution outcomes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.