Fix BENCHMARK_COLS

src/display/utils.py

@@ -179,7 +179,7 @@ TYPES_LITE = [c.type for c in fields(AutoEvalColumn) if c.displayed_by_default a
 EVAL_COLS = [c.name for c in fields(EvalQueueColumn)]
 EVAL_TYPES = [c.type for c in fields(EvalQueueColumn)]
 
-BENCHMARK_COLS = [t.value.col_name for t in Tasks]
+BENCHMARK_COLS = [HEADER_MAP[t.value.col_name] for t in Tasks]
 
 NUMERIC_INTERVALS = {
     "?": pd.Interval(-1, 0, closed="right"),