Inference

Running a trained model to get an answer. This is what happens each time someone sends it a prompt.