Vector embedding
Section titled “Vector embedding”A vector embedding represents an item as a numeric vector so relationships between items can be compared using distance or similarity measures.
Embeddings can represent text, images, users, products, or other objects. Similarity reflects what the embedding model learned, not an absolute semantic truth.
Retrieval systems often use embeddings to find candidate items that are close to a query vector.
Related knowledge
Retrieval-augmented generationRetrieve external knowledge for a task and provide the selected evidence to the model at generation time.