Glossary term
Glossary term
Memory and Retrieval
Multi-vector search improves retrieval by using more than one semantic representation to find relevant information. It helps surface better results by capturing different meanings or perspectives behind a single query.
ColBERT and ColPali use multi-vector representations for late-interaction retrieval.
Vespa, Weaviate, and Qdrant all support multi-vector search through named-vector schemas.
Jina AI offers a ColBERT-style multi-vector embedding model in production deployments.