The gradient of the ReLU function with respect to its input is either 0 or 1. What does this imply for gradient flow in backpropagation?