RAG

RAG at a glance

RAG augments a generative model’s input with retrieved material. Quality depends on retrieval, source freshness, permissions and answer evaluation; RAG does not eliminate all model errors.