When you use a MSE loss function with a softmax output activation function: Select one: Select one: a. it increases the peak gradients of the softmax leading to overfitting b. it reduces the peak gradients of the softmax leading to better learning O C. it reduces the gradients in the tails of the softmax and makes learning difficult d. it smooths the gradients of the softmax overall leading to better learning it increases the gradients in the tails of the softmax and makes learning better e