7 ms·Could a newer language model use this to penalize output that fails the classifier during training?by caxco93 4y agoCould a newer language model use this to penalize output that fails the classifier during training?