[Add] ignore datasets that are not functional atm

src/dataloading.py

@@ -86,6 +86,15 @@ def get_leaderboard_datasets(model_ids):
     if model_datasets:
         common_datasets = set.intersection(*model_datasets.values())
 
+    # Filter datasets that are not MCQ or currently do not work
+    ignore = ["_bbh_", "_gpqa_", "_math_", "_ifeval_"]
+    discard = []
+    for dataset in common_datasets:
+        for ignore_data in ignore:
+            if ignore_data in dataset:
+                discard.append(dataset)
+    common_datasets = [dataset for dataset in common_datasets if dataset not in discard]
+
     return sorted(common_datasets)