Visualizing self attention weights for sequence addition problem with LSTM?

Question

I am using Self Attention layer from here for a simple problem of adding all the numbers in a sequence that come before a delimiter. With training, I expect the