Rotation functions are creating new "actions" and repeating states along a path

Looking at the analysis I just ran, it looks like rotation actions are getting coded as new actions that make new states along a path (the path arrows are looping back through the same state they originate from), and they're printed in the path-data (and I'm assuming also influencing the frequencies and probabilities of each state). Happening for both child and adult data in all 6 tasks.