A hidden layer is a layer in a neural network between the input layer and the output layer.
Hidden layers transform input features into intermediate representations. Multiple hidden layers allow deep-learning models to represent more complex patterns.