← Back to AI Foundations
High School Lab

Neural Net Internals Explorer

Inspect every weight and activation of a small network — then train it with real gradient descent.

Activation Function

σ(z) = 1 / (1 + e⁻ᶻ) — smooth, output in (0, 1), derivative = a(1−a)

Input Values

0.000
1.000

Input → Hidden Weights (W)

-0.591
0.023
0.517
0.071
-0.623
-0.758

Hidden Biases (b_h)

-0.716
-0.712
0.365

Hidden → Output Weights (v)

0.733
0.500
-0.342

Output Bias (b_o)

0.382

Network Diagram

Line thickness = |weight|; cyan = positive, red = negative. Circle fill = activation level.

Live Activations

Inputs: x₁ = 0.000, x₂ = 1.000

Hidden Layer

h1: z = -0.6456a = 0.3440
h2: z = -1.3352a = 0.2083
h3: z = -0.3927a = 0.4031

Output

z_o = 0.5998ŷ = 0.6456

The Math (Forward Pass)

For each hidden neuron j (1..3):

z_j = w[x1→hj]·x1 + w[x2→hj]·x2 + b_hj

a_j = act(z_j)

Output:

z_o = v1·a1 + v2·a2 + v3·a3 + b_o

ŷ = act(z_o)