Use qem for BBH

app.py

@@ -49,7 +49,7 @@ def get_leaderboard_df(merge_values: bool = True):
                 value = data["results"][first_result_key]["acc_norm"]
             # BBH has several metrics but we report just the average one
             elif task.lower() == "bbh":
-                value = [v["em"] for k, v in data["results"].items() if "_average" in k.lower()][0]
+                value = [v["qem"] for k, v in data["results"].items() if "_average" in k.lower()][0]
             else:
                 first_metric_key = next(
                     iter(data["results"][first_result_key])