s3 surpassed static, zero-shot and end-to-end tuned baselines on most benchmarks, achieving strong gains with only 2.4k training examples, significantly less than competitors.
Founders
Patrick (Pengcheng) Jiang
Company Description
s3 is an open-source framework designed to build retrieval-augmented generation (RAG) systems more efficiently, simplifying and reducing the cost of creating retriever models within RAG architectures for developers creating real-world large language model applications.