Definition
The representation of words, passages and entities as points (embeddings) in a high-dimensional space in which distance stands for difference in meaning, so that relatedness can be measured rather than matched on keywords.
How we use the term
Dense retrieval compares a query's vector with the vectors of indexed passages (cosine similarity in a vector database); relevance is a measured closeness.
Also called
Where it sits
Broader
Narrower
No items found.
On this site
Elsewhere
WikidataHW-CON-063
All concepts