Training deeper networks quiz
Question 1 of 5
All weights into a hidden layer of 3 sigmoid neurons start at 0.5. After one backpropagation step, how do the three neurons' gradients compare?
Getting one wrong here is cheaper than getting it wrong in your own code.