🧠 Artificial Neuron Connection System (2D)
Interactive 2D artificial-neuron network: adjust input signals, pick any connection and edit its weight, and watch the forward-pass activation flow live through a configurable hidden layer with sigmoid, tanh, ReLU or step nonlinearities.
This 2D companion turns the network into a live, editable circuit instead of an orbiting decoration: two input sliders feed a configurable hidden layer (2–6 neurons, your choice of tanh, sigmoid, ReLU or step activation), and a connection picker lets you select any single weight in the network — input-to-hidden or hidden-to-output — and drag it from strongly inhibitory to strongly excitatory while small moving dots trace the signal actually flowing along every wire in real time. The live readout panel reports the output neuron's activation, whether it is firing, how many hidden units are active, and the total signal energy passing through the network, so the effect of changing one connection's strength is visible both in the diagram and in the numbers.
Interactive 2D artificial-neuron network: adjust input signals, pick any connection and edit its weight, and watch the forward-pass activation flow live through a configurable hidden layer with sigmoid, tanh, ReLU or step nonlinearities.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install