n 1 2 3 4 01 0.64 0.14 0.39 0.06 22 0.28 0.47 0.34 0.45 93 0.080.39 0.27 0.49 The table shows the predictions for a given choice of weights in softmax regression with 3 states. n corresponds to the index of the data point, (0 n. Yn) where the training labels for the 4 points are 3,1,2,3 for n=1,...,4. Compute the cross-entropy loss. Using the natural logarithm (to the base e) the cross-entropy is 12.71 6.28 15.56 2.86