Laplace Distribution

First saw reading this article https://semianalysis.com/2024/01/11/neural-network-quantization-and-number/

They’re essentially very sharp Gaussian Distributions.

“In practice, neural networks are typically normal or laplace distributed”.

$f(x|\mu ,b)=\frac{1}{2b}\exp \left(-\frac{|x-\mu |}{b}\right),-\infty <x<\infty$