Reading up on retrieval-augmented-generation

An interactive essay explains how agent memory systems work through context windows, embeddings, and retrieval architectures to overcome LLM statelessness.

Using a basic BM25 or e5 retriever with an LLM agent can achieve 0.289→0.453 NDCG on Amazon ESCI by reasoning over queries, but this approach fails on passage retrieval where the embedding model already knows best.