Seen in this image, the AI very obviously inserts comments that it probably should not be inserting. My guess is this is from bad training data. If that is the case, I’m sure that this isn’t an easy fix. However, I thought I should report it anyway. Apologies if this is in the wrong topic.