diff --git a/src/decoupler/bm/_run.py b/src/decoupler/bm/_run.py
index 7f4a35f9..9a330532 100644
--- a/src/decoupler/bm/_run.py
+++ b/src/decoupler/bm/_run.py
@@ -50,7 +50,9 @@ def _tensor_scores(
 
 def _tensor_truth(obs: pd.DataFrame, srcs: np.ndarray) -> pd.DataFrame:
     # Explode nested perturbs and pivot into mat
-    grts = obs.explode("source").pivot(columns="source", values="type_p").notna().astype(float).fillna(0.0)
+    grts = (
+        obs.explode("source").pivot(columns="source", values="type_p").notna().astype(float).fillna(0.0).loc[obs.index]
+    )
     miss_srcs = srcs[~np.isin(srcs, grts.columns)]
     miss_srcs = pd.DataFrame(0, index=grts.index, columns=miss_srcs)
     grts = pd.concat([grts, miss_srcs], axis=1)
diff --git a/tests/bm/test_benchmark.py b/tests/bm/test_benchmark.py
index ae5abd2d..14cf153d 100644
--- a/tests/bm/test_benchmark.py
+++ b/tests/bm/test_benchmark.py
@@ -13,7 +13,7 @@
         [["auc"], None, "expr", False, 0.05, 5, False],
         [["auc", "fscore"], "group", "expr", False, 0.05, 5, False],
         [["auc", "fscore", "qrank"], None, "source", False, 0.05, 2, False],
-        [["auc", "fscore", "qrank"], "group", "source", False, 0.05, 1, False],
+        [["auc", "fscore", "qrank"], "class", "source", False, 0.05, 1, False],
         [["auc", "fscore", "qrank"], "bm_group", "expr", True, 0.05, 5, False],
         [["auc", "fscore", "qrank"], "source", "expr", True, 0.05, 5, False],
     ],