Skip to content
Calcrivo

ELU Activation Calculator

Calculate the ELU (Exponential Linear Unit) activation output for a given input and alpha.

Inputs

Controls the saturation value for large negative inputs. Standard default is 1.0.

ELU(x)

-0.632121

ELU'(x)

0.367879

Step by step

  1. ELU(x) = α × (e^x − 1) (negative branch)

    1 × (e^-1 − 1)

    = -0.632121

  2. Derivative: 1 if x > 0, else ELU(x) + α

    -0.632121 + 1

    = 0.367879

How it works

ELU behaves like the identity for positive inputs but smoothly saturates toward −α for large negative inputs instead of flattening to a fixed slope: ELU(x) = x for x > 0, and α·(e^x − 1) for x ≤ 0. Unlike ReLU and Leaky ReLU, ELU is smooth (continuously differentiable) at x=0, and its negative saturation value pushes mean activations closer to zero across a layer, which can speed up training convergence compared to ReLU-family functions that only allow non-negative (or unboundedly negative, for Leaky ReLU) outputs.

Formula

ELU(x) = x if x > 0, else alpha × (e^x - 1)

x
Input value
alpha
Saturation magnitude for negative inputs (default 1.0)

Frequently Asked Questions

Why does ELU saturate at −α instead of continuing linearly like Leaky ReLU?

The exponential term e^x approaches 0 as x becomes very negative, so α×(e^x − 1) approaches −α — this bounded negative saturation (rather than an unbounded linear decrease) helps push a layer's mean pre-activation closer to zero, which some research suggests improves training dynamics.

Is ELU differentiable everywhere, including at x = 0?

Yes — unlike ReLU and Leaky ReLU which have a sharp corner at x=0, ELU's positive and negative branches meet with matching slopes at x=0 when derivative continuity is checked, making it a smooth (C¹ continuous) function throughout.

What's a typical value for alpha in ELU?

1.0 is the standard default from the original ELU paper and is used in the vast majority of implementations; larger alpha values increase how negative the saturation point can go, but 1.0 works well across most architectures without additional tuning.

You might also need