ModelRefs / Self-RAG — AI Glossary
Self-RAG — AI Glossary
A RAG approach where the model generates special reflection tokens to decide whether to retrieve, assess retrieved chunks, and critique its own output.
Overview
Self-RAG (Asai et al. 2023) fine-tunes the generator model to emit four reflection token types: [Retrieve] (retrieval needed?), [IsREL] (is this chunk relevant?), [IsSUP] (is the output supported?), [IsUSE] (is the output useful?). Enables dynamic, selective retrieval and self-critique without a separate orchestrator.
Reference details
| Topic | rag |
|---|---|
| Last reviewed | 2026-06-24 |
Related terms
Commonly confused with
A specific trained mechanism, not a prompting pattern: the model emits reflection tokens deciding whether retrieval is needed and critiquing what came back. That distinguishes it from agentic RAG, which describes the same behaviour achieved with any model and a scaffold, and from Corrective RAG, which grades documents and falls back to another source rather than reflecting on its own output.
Primary source
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Self-RAG — AI Glossary.
Frequently asked questions
What is Self-RAG?
A RAG approach where the model generates special reflection tokens to decide whether to retrieve, assess retrieved chunks, and critique its own output.
What concepts are related to Self-RAG?
Closely related concepts include corrective rag, agentic rag, retrieval pipeline.