The proportions for a document sum to 1.
The word probabilities shown here are truncated to the top terms. In a fitted model the full topic distribution spans the vocabulary.