It’s certainly intentional. It’s certainly possible to train the model not to respond that way.