Retrieval-augmented generation (RAG)

Retrieval-augmented generation (RAG) is a generative AI technique that combines relevant content retrieval and content generation to enhance the output of a generative large language model (LLM). By grounding the LLM to the retrieved content, it allows the LLM to generate content that’s more accurate, contextually relevant, and up to date.

Relevance Generative Answering (RGA) uses Retrieval-augmented generation (RAG) to generate answers to user queries. A Passage Retrieval (CPR) model retrieves the passages for your enterprise’s Retrieval-augmented generation (RAG) system to enhance your large language model (LLM)-powered applications.