diff --git a/torchrl/_utils.py b/torchrl/_utils.py
index 56f5c32b3af..cfc54ad207c 100644
--- a/torchrl/_utils.py
+++ b/torchrl/_utils.py
@@ -885,7 +885,7 @@ def _standardize(
 
     Args:
         input (Tensor): the input tensor to be standardized.
-        exclude_dims (Sequence[int]): dimensions to exclude from the statistics, can be negative. Default: ().
+        exclude_dims (Tuple[int]): dimensions to exclude from the statistics, can be negative. Default: ().
         mean (Tensor): a mean to be used for standardization. Must be of shape broadcastable to input. Default: None.
         std (Tensor): a standard deviation to be used for standardization. Must be of shape broadcastable to input. Default: None.
         eps (float): epsilon to be used for numerical stability. Default: float32 resolution.