diff --git a/fastchat/llm_judge/gen_model_answer.py b/fastchat/llm_judge/gen_model_answer.py
index 9d6995227..c2f73029d 100644
--- a/fastchat/llm_judge/gen_model_answer.py
+++ b/fastchat/llm_judge/gen_model_answer.py
@@ -125,7 +125,7 @@ def get_model_answers(
                         output_ids,
                         spaces_between_special_tokens=False,
                     )
-                    if conv.stop_str:
+                    if conv.stop_str and output.find(conv.stop_str) > 0:
                         output = output[: output.find(conv.stop_str)]
                     for special_token in tokenizer.special_tokens_map.values():
                         if isinstance(special_token, list):