Sparse Vectors
The keyword-tracking mathematical representation used in lexical search — why precise terminology on your pages still matters.
Plain Language Definition
Sparse vectors are word lists that track which specific keywords appear in a document, with most slots empty because most words do not appear in any single document. They are the mathematical backbone of keyword search systems like BM25. Think of them as a very long checkboxes-of-words sheet where most boxes are unchecked.
Technical Definition
High-dimensional vector representations where dimensions map directly to vocabulary entries, with non-zero values only for terms actually present in the document — enabling efficient inverted-index lookup in lexical retrieval systems with minimal computation.
Why This Matters for AI Search Visibility
Sparse vector retrieval excels at finding exact keyword matches — which is why using precise technical terminology on your pages still matters. Hybrid retrieval systems combine sparse (keyword) and dense (semantic) vectors, so content needs to perform well in both dimensions.