RAG and Search — Chunking · Embedding · Reranking · Hybrid
RAG (Retrieval-Augmented Generation) is an approach to supplement an LLM's limited training cutoff and memory limits with external knowledge. By separating retrieval and generation, you can handle new material without r…