Finding similarity between sets of predictions distribution vectors

Question

I have a LDA classification model that identifies topics in a given text. The output of the model for a text is a list distribution of probability factors where each factor