diff --git a/nodes.py b/nodes.py
index 35fe7e3..42f7013 100644
--- a/nodes.py
+++ b/nodes.py
@@ -299,7 +299,7 @@ class CogVideoTextEncodeCombine:
         elif combination_mode == "weighted_average":
             embeds = conditioning_1 * (1 - weighted_average_ratio) + conditioning_2 * weighted_average_ratio
         elif combination_mode == "concatenate":
-            embeds = torch.cat((conditioning_1, conditioning_2), dim=-1)
+            embeds = torch.cat((conditioning_1, conditioning_2), dim=-2)
         else:
             raise ValueError("Invalid combination mode")