In the first part of this exercise, you train a network on a version of the first monks problem, looking for a particular pattern of banding in the network's hidden units. In the second part of the exercise, you examine some statistics that have already been performed for you on such bands in an attempt to interpret features that the hidden units have represented locally.