Softmax Temperature Visualization

What is Softmax?
Softmax transforms a vector of numbers into a probability distribution. Each output is between 0 and 1, and all outputs sum to 1.
Temperature Effect:
As you increase the temperature multiplier (scale inputs up), the softmax becomes more "confident" - it approaches a step function where the highest input gets probability ≈1 and others get ≈0. This is what people mean by "acting like a step function"!
Low (0.1x) - Smooth, uniform High (10x) - Sharp, step-like
softmax(xi × 1.0) = exp(xi × 1.0) / Σ exp(xj × 1.0)

Softmax Outputs (Probabilities)

Class 1
0.33
Class 2
0.33
Class 3
0.33
Sum of probabilities: 1.000