Anatomy

Diagrams of boxes and arrows describe an idea. These build it. Each piece is the real thing, running: a transformer you can type into, laid out in 3D from its own numbers; a web stack you can overload. The 3D dissections walk through a landmark paper or system component by component, trace every figure to its source, and end with experiments of their own.

11 models 5 AI & ML5 Architecture1 Mechanistic Interpretability 943 parts and components 270 links into the libraries

All11 AI & ML5 Architecture5 Mechanistic Interpretability1

Any Language models 2 Neural network foundations 2 Training & alignment 1

Any Working model 5

1 model for “adam”.

Working model · 3D AI & ML · Training & alignment

Anatomy of Adam

Kingma and Ba's optimiser, dissected: every line of Algorithm 1 drawn as real blocks on a 151-parameter network, checked against PyTorch, with a research desk that reruns Reddi et al.'s counterexample, finds how rare a large gradient must be to defeat it, and tests the paper's claims about bias correction, the size of the step and ε.

Adam in 3D: its gradient, two moving averages, bias corrections and step drawn as real blocks for any step of a live training run, tested against PyTorch, then a pre-registered research desk on the c…

61 components · Intermediate 18 Oct 2026