06
Vector stores and scores
Where vectors live, what a similarity score means, and many vectors per chunk.
Concept 6.1
Vector databases, ANN indexes and choosing a store
When exact search is enough, what approximate indexes like HNSW trade away, and how compression shrinks vectors 32 times, with measured recall on our corpus.
Concept 6.2
Score calibration and thresholds
Can a similarity cut-off tell "no good match" from a real answer? We scored 47 answerable and 15 unanswerable questions; the ranges overlap, so a threshold is one guard, not the guard.
Concept 6.3
Late interaction and multivector retrieval
Instead of one vector per chunk, keep many (per sentence, or per token as in ColBERT) and score by the best match. We tested the sentence version; it used 12 times the vectors for the same result.
Chapter 6 recap