Layers and networks

Neurons are stacked in layers, and each layer's outputs become the next layer's inputs. A whole layer is one matrix multiplication followed by an activation. Early layers pick up simple patterns and later ones combine them into complex ones. "Deep" learning just means many layers; large models have billions of weights.

Why it matters: Language models are very large networks, so this is the shape of the machine.

Builds on: A single neuron, Matrices and matrix multiplication

Learn it, then prove it

Sign in (free) to take this step's mastery check and save your journey.