Softmax with Temperature

A C# console application demonstrating how the softmax function behaves with different temperature values.

The program takes an array of source logits and computes the probability distribution for each temperature from a given set.

It visually illustrates how temperature affects the "sharpness" of the distribution: when T < 1, differences between logits are amplified; when T > 1, they are smoothed out.

Example output

  Source |    0.5 |    1.0 |   10.0 |
--------------------------------------
    0.00 |   5.06 |  15.00 |  31.24 |
    1.39 |  80.90 |  60.00 |  35.88 |
    0.51 |  14.04 |  25.00 |  32.88 |