what attribute mean in dataset, positive and label

[Miraclemarvel55]

In the probe subset , what mean positive and label if true or false and original_human_prediction": "2",
"flipped_human_prediction" what the number mean ,please share the definition, tks.

[SiyuanWangw]

positive in the probing subset refers to the presence or absence of negation words in the conclusion of a rule. Specifically, positive=True indicates that there are no negation words (such as CanDrive) whereas positive=True signifies the inclusion of negation words (such as CanNotDrive).

label means whether this rule is verified by human expert as original_human_prediction="2"(True/Right/Yes) and flipped_human_prediction="1"(False/Wrong/No). The definition of original probing and flipped probing can be found in our paper (the second paragraph of Section 3.1).

[Miraclemarvel55]

Really appreciate your work.