A model whose output is vectors rather than text, used to power semantic search, recommendations and similarity matching.
Also called text-embedding-3.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Embeddings model in context, with comparison tables and the common traps.
Related terms
- External model
A database object, made with CREATE EXTERNAL MODEL, that stores the details AI_GENERATE_EMBEDDINGS needs to call an AI model: endpoint, API format, model type (EMBEDDINGS), model name and credential. It is changed with ALTER, and dropping it leaves the credential in place.