Definition
A linear architectural pattern where an LLM retrieves relevant document chunks from a vector database based on semantic similarity to a query and uses that retrieved context to generate a response. It prioritizes a direct 'Retrieve-then-Read' flow without iterative refinement or complex query transformations.
Distinguish from Advanced RAG; Basic RAG lacks pre-processing (query expansion) and post-processing (re-ranking) steps.
"An open-book exam where the student looks up a single highlighted passage in a textbook before writing their answer."
- Vector Database(Component)
- Embedding Model(Prerequisite)
- Naive RAG(Synonym)
- Advanced RAG(Architectural Evolution)
Conceptual Overview
A linear architectural pattern where an LLM retrieves relevant document chunks from a vector database based on semantic similarity to a query and uses that retrieved context to generate a response. It prioritizes a direct 'Retrieve-then-Read' flow without iterative refinement or complex query transformations.
Disambiguation
Distinguish from Advanced RAG; Basic RAG lacks pre-processing (query expansion) and post-processing (re-ranking) steps.
Visual Analog
An open-book exam where the student looks up a single highlighted passage in a textbook before writing their answer.