Or you get the original output result was perfect and the adversarial "rethinking" switches to an incorrect result.
this seems to happen far more than i would like
this seems to happen far more than i would like