Fig.1

Concept

Multi-head attention

Multi-head attention runs several attention operations in parallel, each with its own learned projections, then concatenates their outputs. It generalizes the single scaled dot-product attention introduced in "Attention Is All You Need," which computes `softmax(QK^T /…

The rest of “Multi-head attention” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.

Log in to unlock

← Back to the library

Multi-head attention, explained · Fig. 1