So the probabilities would be skewed towards something, but unless the model could somehow count/infer its own weights, I don't see how it could "introspect" to

So the probabilities would be skewed towards something, but unless the model could somehow count/infer its own weights, I don't see how it could "introspect" to see if something is more common than something else.