AI Research
Vectorless RAG Hits 98.7% on FinanceBench
A new approach to Retrieval Augmented Generation (RAG) is making waves. It ditches vector embeddings entirely, achieving near-perfect results on a key benchmark.