ClassroomThe classroom
Back to the lesson libraryMECHANISM · 11 MIN
M26.5 CONNECT THE MECHANISM

Retrieve evidence and condition an answer on it

Give a language model an open-book exam. Follow one customer question from 20,000 chunks to a cited answer, and learn to find the exact step where a wrong answer went wrong.

LESSON OVERVIEW11 min lesson

Lesson overview

Give a language model an open-book exam. Follow one customer question from 20,000 chunks to a cited answer, and learn to find the exact step where a wrong answer went wrong.

What you’ll explore

  • A RAG system retrieves external material and supplies selected evidence to a generator; ingestion, retrieval, context construction, generation, and verification each contribute distinct failure modes.
GO TO THE SOURCE

Original explanations, connected to the research.

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (Lewis et al., 2020)Lost in the Middle: How Language Models Use Long Contexts (Liu et al., 2023)Introduction to Information Retrieval — authors’ online edition
Suggest a correction

A precise note can make an explanation better.

Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.

The file includes this note, the scene title, and lesson metadata. Your saved progress and quiz responses are excluded. Download before leaving or reloading to keep your note.