What is the core mechanism that distinguishes Self-RAG (Asai et al. 2023) and Corrective-RAG (CRAG, 2024) from a standard single-shot retrieve then generate pipeline?

MCQ·Hard·4.0 · 0·~1 min·Relevant atAnthropic Databricks LangChain

Attempt it

Quick reference

TL;DR

Both add a self-reflective quality check on retrieval and act on it: Self-RAG via reflection tokens, CRAG via a lightweight evaluator. Standard RAG just trusts whatever came back.

Memory aid

Easy to grasp

Imagine you ask a friend to look something up in a library. Standard RAG is a friend who grabs the first books off the shelf and reads from them no matter what. Self-RAG and CRAG are friends who pause and ask themselves: are these books actually about the question? If not, they go back, try a different shelf, or search online. Self-RAG was trained to spend extra effort emitting little check mark words as it works, like 'do I even need a book here?' and 'does the book actually answer this?' CRAG is simpler: it has a small helper that scores how relevant the books look, and if the score is low, it switches strategies. Both share one big idea: do not blindly trust the first results.

Key concepts

RAG system design Embeddings Vector databases

Detailed answer & concept explanation

~7 min read

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example. Click to expand.

Everything you need to truly understand this topic: intuition, mechanics, step by step explanation, code, formulas, and worked example.

Quick scan · 2–4 min

Everything important, quickly.

8 min: explain the shared insight, walk Self-RAG's four reflection tokens, walk CRAG's three-grade evaluator, contrast with multi-query, and close with the 2026 tool-use reproduction.

Code example

python

# CRAG-style corrective gate.
from enum import Enum

class Grade(Enum):
    CORRECT = "correct"
    AMBIGUOUS = "ambiguous"
    INCORRECT = "incorrect"

async def crag_answer(query: str) -> str:
    chunks = await retrieve(query, top_k=8)
    grade = await evaluator.grade(query, chunks)  # T5-style classifier

    if grade == Grade.CORRECT:
        context = await knowledge_refinement(chunks)  # strip level filter
    elif grade == Grade.INCORRECT:
        rewritten = await rewrite_query(query)
        context = await web_search(rewritten)
    else:  # AMBIGUOUS
        web = await web_search(query)
        context = await knowledge_refinement(chunks) + web

    return await generator.complete(query, context)

Side-by-side

Aspect	Standard RAG	Self-RAG	CRAG
Quality gate	none	reflection tokens in generator	external evaluator model
Decides when to retrieve	always retrieves	[Retrieve] token per segment	always retrieves, then grades
Corrective action	none	self-critique, skip, refuse	web search, query rewrite
Generator requirement	any frozen LLM	fine-tuned on reflection data	any frozen LLM
Extra inference cost	1x	1.2-1.5x (reflection tokens)	1.1-1.3x (one evaluator call)
2026 adoption	default everywhere	via tool-use on frontier models	common in LangChain / LlamaIndex

Real-world examples

Real products, models, and research that use this idea.

Self-RAG paper by Asai et al. ships a 7B and 13B fine-tuned LLaMA-2 model that emits the four reflection tokens during decoding.
CRAG paper by Yan et al. uses a T5-large evaluator and falls back to Google Search via the Serper API for the corrective web step.
LangChain's adaptive rag template implements a CRAG-style gate using a grader then search router.
LlamaIndex ships a Corrective RAG query engine that wraps any retriever with a relevance score then decide gate.
Anthropic Claude 4.7 contextual retrieval combined with tool-use achieves Self-RAG-like gating without fine-tuning: the model decides when to call the retrieve_documents tool.

Interviewer follow-ups

What an interviewer would ask next. Try answering before peeking at the approach.

QHow would you implement Self-RAG-style gating on top of a closed weight frontier model like Claude Opus 4.7 or GPT-5.5?

QCRAG's evaluator is one extra inference call per query. How would you train it, and what is the latency budget?

QCompare the failure modes of Self-RAG and CRAG when the corpus and the question are both out of distribution.

QHow does Adaptive RAG combine Self-RAG / CRAG patterns with multi-hop decomposition?

Don't say this

Red flags and common mistakes that signal junior thinking. Click to expand.

Most common mistake

Conflating these with multi-query retrieval. Multi-query expands the question into paraphrases and runs them in parallel; Self-RAG and CRAG inspect the results and react to quality.

60-second cheat sheet

The night-before-the-interview bullets. Scan these on the way to the call.

References

Primary sources. Skim if you want the original framing.