Loading 3D engine…
🧠 Neural Network — 3D forward pass
Neurons: 0 · Connections: 0 · Parameters: 0
Each neuron computes y = f(Σ wᵢxᵢ + b) — the same weighted-sum-plus-bias formula, with a real activation function f, laid out layer by layer along the X axis in true 3D space. Weight sign colors each connection (blue = positive, red = negative); the pulse shows the signal actually propagating forward, layer by layer.
Drag to orbit · Scroll to zoom · Rebuild changes the architecture, forward pass runs the real math