hard20 pts · first answer countsvector-databasesrag

Retrieval got worse as the index grew

An HNSW vector index grew from 1 million to 50 million chunks with the same settings. Answer quality dropped, and recall@10 against exact search fell from 0.97 to 0.85. What should you tune first?

A.Lower the embedding dimension to 2B.Raise the search-time candidate list (ef_search), and if needed rebuild with a higher M, measuring recall and latencyC.Switch from cosine to random rankingD.Shrink the number of chunks returned to 1

← Error budget for 99.9%User asks for data deletion →