Whether a model applies the same reasoning strategy when classifying different instances of the same category.