Dot product
Multiply two lists of numbers pairwise and add it all up — one number that says how aligned they are.
Think of it like
Scoring how well two playlists match: multiply overlapping tastes, sum the score, higher means more alike.
Example
Attention scores “how much should token A care about token B” with a dot product of their vectors.
How it actually works
The dot product of two vectors sums the products of matching components. It’s the atom of matrix multiplication and the cheapest measure of similarity — big and positive when vectors point the same way, near zero when unrelated. Normalize it and you get cosine similarity.
For product teams
It’s the tiny operation underneath “how similar are these two things” across search and attention.
For engineers
⟨a,b⟩ = Σ aᵢbᵢ; the elementary op composing matmul and the unnormalized similarity in attention.
Related
- Matrix multiplication — Composed into matrix multiplication.
- Cosine similarity — Normalized becomes cosine similarity.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome