It does not work with the extractors because you used the preprocessor (ColumnTransformer) to fit and transform. You can get them by specifying the step within the ColumnTransformer:

preprocessor["cat"].get_feature_names_out()