Retrieval augmented generation (RAG) encompasses a family of systems that extend conventional language models , large and otherwise (LLMs), to incorporate context based on retrieved knowledge from a document base, thereby leading to more truthful and relevant responses being generated upon user queries.



Source link