Some comments on the connection between disordered long range spin glass models and their mean field version