add

app.py

@@ -62,7 +62,7 @@ def init_leaderboard(dataframe):
     if dataframe is None or dataframe.empty:
         raise ValueError("Leaderboard DataFrame is empty or None.")
 
-    # pdb.set_trace()
+    pdb.set_trace()
     num_cols = dataframe.select_dtypes(include=['number']).columns
     styler = dataframe.style.format({col: "{:.1f}" for col in num_cols})
     return gr.components.Dataframe(

src/leaderboard/read_evals.py

@@ -87,7 +87,7 @@ class EvalResult:
         for metric in Metrics:
             metric = metric.value
 
-            results[metric.short] = data["results"][task][metric.short]
+            results[metric.short] = float(data["results"][task][metric.short])
 
         return self(
             eval_name=result_key,