🧠 Neural Network Visualiser
Operation
AND
OR
XOR
Hidden neurons
2
3
Learning rate
🎲 Randomise weights
Network
positive weight
negative weight
forward signal
backprop (δ) signal
line thickness = |weight|
What the network is doing
Click a “Train” button to watch the forward pass, error, and weight updates step by step.
Training data
x₁
x₂
target
prediction
Animation
▶ Train one full pass
⏩ Train 100 epochs (no anim)
Total epochs trained:
0
Status
Mean squared error
—
Accuracy (rounded)
—
Use the network
x₁
x₂
Run →
—