HyDE
Also known as: Hypothetical Document Embeddings
A query-expansion technique for retrieval-augmented generation in which an LLM is first asked to generate a hypothetical answer to the user's question, and that hypothetical answer — rather than (or alongside) the raw query — is embedded and used to search the document index. HyDE improves recall on short, ambiguous, or jargon-laden queries where the original wording may not match the documentation's vocabulary. It increases latency and cost, so in accessibility systems like AskEase it is typically offered as an optional accuracy-over-speed mode.
Category: AI · technology
Related: RAG · Context Engineering