Empirical Comparisons of MASC Word Sense Annotations