A Note on Saturated Models for Many-Valued Logics