Which pgvector distance metric is equivalent to maximizing inner product (used with unit-normalized embeddings)?

1. A <-> (Euclidean)
2. B <=> (Cosine)
3. ✓ C <#> (Negative inner product) (correct answer)
4. D <@> (L1)

Shortcut: <#> returns negative inner product. Sorting ASC by <#> finds vectors with maximum dot product. For unit-normalized embeddings, inner product and cosine similarity are equivalent.