HyDE
Embedding a generated answer instead of the question.
Overview
HyDE asks the model to write a hypothetical answer to the question, then embeds that answer and searches with it. It solves the problem that a question and its answer share fewer terms and less structure than two answers do, so the query embedding lands far from the passages that would answer it. It exists as an inversion that works precisely because the generated answer need not be correct to be useful.
Technology Graph
How HyDE connects to the rest of the stack.
Uses and depends on
- Embedding concept Techniques HyDE uses Embedding High ✓ source
Applied to
- RAG technique Techniques HyDE used for RAG High ✓ source