diff --git a/eval/benchs/halueval/eval_halueval_dialog.py b/eval/benchs/halueval/eval_halueval_dialog.py
index 8cf0110..45968fd 100644
--- a/eval/benchs/halueval/eval_halueval_dialog.py
+++ b/eval/benchs/halueval/eval_halueval_dialog.py
@@ -51,7 +51,7 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""
diff --git a/eval/benchs/halueval/eval_halueval_qa.py b/eval/benchs/halueval/eval_halueval_qa.py
index 6258eea..4476424 100644
--- a/eval/benchs/halueval/eval_halueval_qa.py
+++ b/eval/benchs/halueval/eval_halueval_qa.py
@@ -56,14 +56,13 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""
         # Remove the leading "#", ".", ","
         answer = answer.strip("#").strip(".").strip(",")
 
-
         return {
             "metrics": {
                 "correct": ground_truth.lower() == answer.lower(),
diff --git a/eval/benchs/halueval/eval_halueval_summa.py b/eval/benchs/halueval/eval_halueval_summa.py
index 7f2201f..f974470 100644
--- a/eval/benchs/halueval/eval_halueval_summa.py
+++ b/eval/benchs/halueval/eval_halueval_summa.py
@@ -42,7 +42,7 @@ def scoring(self, data_point: dict) -> dict:
         )
         response = self.model.safe_request(query)
 
-        answer = response.strip().split()  
+        answer = response.strip().split()
         # Extract the first word, such as "Yes", "No", "#Yes", "No."
         # Note: "".strip() returns [] instead of [""]
         answer = answer[0] if answer else ""
diff --git a/eval/benchs/uhgeval/eval_sele.py b/eval/benchs/uhgeval/eval_sele.py
index 80a4da4..12f258f 100644
--- a/eval/benchs/uhgeval/eval_sele.py
+++ b/eval/benchs/uhgeval/eval_sele.py
@@ -41,8 +41,8 @@ def which_is_true(self, contn1: str, contn2: str, obj: dict) -> tuple[int, str]:
         """Given two continuations, determine which one is more accurate.
 
         Returns:
-            tuple: (answer, response). `answer` is 1 if the first continuation is more 
-            accurate, 2 if the second one is more accurate, and -1 if error. `response` 
+            tuple: (answer, response). `answer` is 1 if the first continuation is more
+            accurate, 2 if the second one is more accurate, and -1 if error. `response`
             is the model's response.
         """
         query = PROMPT_TEMPLATE.format(