Dot-product similarity
The idea: Dot product / cosine as an alignment score: multiply-and-add.
What you'll be able to do: You can explain how a model measures meaning with a dot product.
The problem it solves: How does the machine measure how near two vectors are?
Builds on: Embeddings: meaning as coordinates
← Embeddings: meaning as coordinates · Next: Below the vector: tokens →
All lessons