---
title: "Discussion on \"How RAG Grew Up, Part 1: The Open-Book Exam\" | Hashnode"
description: "Discussion on \"How RAG Grew Up, Part 1: The Open-Book Exam\". Part 1 of 3. This one covers 2017 to 2020, when search and text generation ended up inside the same model."
url: "https://hashnode.com/posts/how-rag-grew-up-part-1-the-open-book-exam/6afa52e8cbdc628a8c23e951"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [How RAG Grew Up, Part 1: The Open-Book Exam](https://1tushar.hashnode.dev/how-rag-grew-up-part-1-the-open-book-exam?utm_source=hashnode&utm_medium=feed)

Part 1 of 3. This one covers 2017 to 2020, when search and text generation ended up inside the same model.

1d ago · 8 min read · 1tushar.hashnode.dev  
Read the article

-   [#artificial-intelligence](https://hashnode.com/tag/artificial-intelligence)
-   [#machine-learning](https://hashnode.com/tag/machine-learning)
-   [#nlp](https://hashnode.com/tag/nlp)
-   [#rag](https://hashnode.com/tag/rag)

## Written by

[Tushar Verma](https://hashnode.com/@1tushar)

@1tushar

  
[View profile](https://hashnode.com/@1tushar)

Data Science Enthusiast

## Responses 1

Markdown supported · ⌘↵ to post

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[18h ago](https://hashnode.com/posts/how-rag-grew-up-part-1-the-open-book-exam/6afa52e8cbdc628a8c23e951/comment/6afb73df4523ddb00ebb7b1d)

Training the query encoder from answer supervision is the historical detail I would carry into the next part. It makes passage usefulness observable through the answer objective, but an answer string appearing in a passage can still be attached to the wrong entity or relation.

For the hostel-fee example, a useful control would include two notices with the same date for different fees. Checking which notice gets retrieved and which source supports the answer would show why retrieval training and source attribution remained separate engineering problems as these systems evolved.
