DocumentationArchitecture

OutputHead

A zoomed-in view of one operation in Piro’s stateful inference loop.

Architecture detail

OutputHead

The baseline output head reads runtime weights; a CTM variant may read recurrent state instead.

← Full model
outputₖ = OutputHead(hₖ₊₁)

What it does

Reads the current runtime weights into the externally returned output.

Why it exists

The baseline needs a stable output boundary after each adapted forward pass.

← Back to architecture