Indeed. They do not mention a "collective increase", they merely state that the climb paused and then continued. There is implication that the pause was coincident with pedal movements and associated (correct) yaw which then became yaw that did not reflect the pedal demand.
Head scratch moment - if the tail rotor demand reduces, the system sends more to the main rotor. Remember the collective/pedal interaction?