plp34 commited on
Commit
795529c
1 Parent(s): cdb5c3d

Correct description for label

Browse files

"Entail" means "to have as a logically necessary consequence." The original description for label got it backwards. The label "entail" means the premise entails the hypothesis, not the other way around; it's also the case for the other two labels. Consider this example:

Premise: A few people in a restaurant setting, one of them is drinking orange juice.
Hypothesis: The diners are at a restaurant.
Label: entailment

Obviously if you know the premise that "a few people [are] in a restaurant, one of them is drinking orange juice", it "has as a logically necessary consequence that "the diners are at a restaurant." However, if you know the hypothesis that they are at a restaurant, "one of them is drinking orange juice" is not its logically necessary consequence.

Please correct this in order not to mislead people on this model card.

Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -131,7 +131,7 @@ The average token count for the premises and hypotheses are given below:
131
 
132
  - `premise`: a string used to determine the truthfulness of the hypothesis
133
  - `hypothesis`: a string that may be true, false, or whose truth conditions may not be knowable when compared to the premise
134
- - `label`: an integer whose value may be either _0_, indicating that the hypothesis entails the premise, _1_, indicating that the premise and hypothesis neither entail nor contradict each other, or _2_, indicating that the hypothesis contradicts the premise. Dataset instances which don't have any gold label are marked with -1 label. Make sure you filter them before starting the training using `datasets.Dataset.filter`.
135
 
136
 
137
  ### Data Splits
 
131
 
132
  - `premise`: a string used to determine the truthfulness of the hypothesis
133
  - `hypothesis`: a string that may be true, false, or whose truth conditions may not be knowable when compared to the premise
134
+ - `label`: an integer whose value may be either _0_, indicating that the premise entails the hypothesis, _1_, indicating that the premise neither entails nor contradicts the hypothesis, or _2_, indicating that the premise contradicts the hypothesis. Dataset instances which don't have any gold label are marked with -1 label. Make sure you filter them before starting the training using `datasets.Dataset.filter`.
135
 
136
 
137
  ### Data Splits