>benparr .

> @openAI has put in a lot of work into RAG (retrieval augmented generation).

>

> Accuracy has jumped from 45% to 98% with things like retaining, prompt engineering, and query expansion.

>

>experilearning What's the context for this slide, is the full presentation available anywhere?