Reproduce “RAILS: Retrieval-Augmented Incremental LLM Clustering at Scale” 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
Armin Oliya, Aleksandra Sawczuk, Radosław Białobrzeski
RAILS clusters ongoing text streams by retrieving likely labels from a persistent pool and asking an LLM to assign tickets in batches, reuse or create labels, or mark noise. Embeddings guide retrieval and label merging, while the LLM decides assignments. Across six public benchmarks, RAILS with Sonnet 4.5 averages 59.3% accuracy, versus 51.2% for LLM-MemCluster; production results use an LLM-labelled reference, not human ground truth.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.