Classic non-linear split
MediumXOR compact
Shows why a linear boundary is not enough. A tiny hidden network learns quadrant-based separation.
Experiment gallery
Each example is designed to expose a different learning behavior: non-linear boundaries, convergence speed, noisy errors, gradient flow, or custom CSV training.
Search examples
Classic non-linear split
MediumShows why a linear boundary is not enough. A tiny hidden network learns quadrant-based separation.
Fast convergence demo
EasyA strong demo for decision boundaries. The network learns an enclosed class region quickly.
Process-view stress test
HardGreat for inspecting confidence, misclassifications, gradients, and the limits of tiny networks.
One boundary sanity check
EasyProves the system works on linearly separable data before trying harder shapes.
Bring your own data
VariableTurns Axon from a playground into a real local experiment lab for 2D binary datasets.