diff --git a/gym/decision_transformer/models/decision_transformer.py b/gym/decision_transformer/models/decision_transformer.py
index f76f8f3b..2b63722e 100644
--- a/gym/decision_transformer/models/decision_transformer.py
+++ b/gym/decision_transformer/models/decision_transformer.py
@@ -134,7 +134,7 @@ def get_action(self, states, actions, rewards, returns_to_go, timesteps, **kwarg
         else:
             attention_mask = None
 
-        _, action_preds, return_preds = self.forward(
+        _, action_preds, _ = self.forward(
             states, actions, None, returns_to_go, timesteps, attention_mask=attention_mask, **kwargs)
 
         return action_preds[0,-1]