The sample data follows the CoNLL-2006 format. Each column  is explained as follows:

0: word_id in the sentence
1: word
2:
3:
4: head_id according to our guideline, -1 if not annotated
5: label according to our guideline, none if not annotated
6: head_id according to HIT-CDT guideline
7: label according to HIT-CDT guideline
8:
9: 

