i seem to have the reverse question everybody usually has regarding CatBoost Regression. My model extrapolates - which i thought was not possible and is in this case counterprod
