ZH_DE_Datasets Format Each sample contains only one current predicate. 1st column: sample id 2nd column: words in sample 3rd column: 'Y' indicates which word is predicate 4th column: sense id of the predicate 5th column: semantic roles of each word