I assume it would be consistent to have this as well in the output of `ml_estimation` and `rank_regression`