diff --git a/docs/source/advanced/using-graphbolt.rst b/docs/source/advanced/using-graphbolt.rst
index 21e3fe516..177e95be0 100644
--- a/docs/source/advanced/using-graphbolt.rst
+++ b/docs/source/advanced/using-graphbolt.rst
@@ -186,7 +186,7 @@ the partitioned DGL graph files have been created on S3.
 After running your distributed partition SageMaker job as normal using
 ``sagemaker/launch_partition.py``, you next need to launch the
 ``sagemaker/launch_graphbolt_convert.py`` script, passing as input
-the S3 URI where the DistDGL partition data is stored by ``launch_partition.py``,
+the S3 URI, where the DistDGL partition data is stored by ``launch_partition.py``,
 **plus the suffix `dist_graph`** as that's where GSPartition creates the partition files.
 
 For example, if you used ``--output-data-s3 s3://my-bucket/my-part-graph`` for
diff --git a/sagemaker/launch/common_parser.py b/sagemaker/launch/common_parser.py
index ce88e6101..d7caaf57e 100644
--- a/sagemaker/launch/common_parser.py
+++ b/sagemaker/launch/common_parser.py
@@ -95,7 +95,7 @@ def parse_unknown_gs_args(unknown_args: List[str]) -> Dict[str, str]:
 
     The input is a list of arguments, the second element of the tuple
     returned by ``argparse.ArgumentParser.parse_known_args()``, which
-    can be a single string depending on how the arguments were passed.
+    can be a single string depending on how the arguments are passed.
 
     We must handle cases like
         ``--target-etype query,clicks,asin query,search,asin``