Update README.md

README.md

@@ -30,6 +30,7 @@ mapping = {"NOT VERIFIABLE": 0, "VERIFIABLE": 1}
 Finally, a consistency check with the labels reported in the original FEVER dataset is performed.
 
 NOTE: no label is available for the "test" split.
+NOTE: there are 3 instances in common between `dev` and `train` splits.
 
 
 ## Code to generate the dataset
@@ -177,4 +178,20 @@ if __name__ == "__main__":
 
     nli_fever = DatasetDict(dataset_splits)
     nli_fever.push_to_hub("pietrolesci/nli_fever", token="<your token>")
+    
+    # check overlap between splits
+    from itertools import combinations
+    for i, j in combinations(dataset_splits.keys(), 2):
+        print(
+            f"{i} - {j}: ",
+            pd.merge(
+                dataset_splits[i].to_pandas(), 
+                dataset_splits[j].to_pandas(), 
+                on=["premise", "hypothesis", "label"], 
+                how="inner",
+            ).shape[0],
+        )
+    #> train - dev:  3
+    #> train - test:  0
+    #> dev - test:  0
 ```