The lecture notes PDF file for course 2 week 3 don’t contain any of Andrew’s scribbles from the video lecture on Softmax (e.g. the activation function).
Also, the section on “Training a Softmax Classifier” (e.g. the loss function, backprop for layer, a^L) is missing entirely from the lecture notes.
Essentially, the lecture notes are devoid of any useful information on Softmax.