Skip to content



A bounded S-shaped function (sometimes called the Logistic function) with an output value between 0 and 1. The output of the sigmoid function has the advantage of being interpretable as a probability, however it is not zero-centered and tends to saturate if inputs become large.

\[ {\displaystyle Sigmoid = {\frac {1}{1+e^{-x}}}} \]


This activation function does not have any parameters.


use Rubix\ML\NeuralNet\ActivationFunctions\Sigmoid;

$activationFunction = new Sigmoid();