Query, key, value

The three vectors an attention layer produces from three linear layers. The key and query are multiplied to form an attention filter (a correlation matrix of how words relate), and that filter then weights the value. The recipe is the scaled dot-product: multiply key and query, divide by the square root of the hidden dimension, then softmax.

Terms like this come up in real clinical scenarios across the HelloAI courses: bite-sized modules with verifiable certificates. An account takes one minute, no password needed.