[CI] Reorganize scoring tests (vllm-project#38207)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>

Author: noooop

tests/entrypoints/pooling/basic/test_truncation.py

@@ -85,7 +85,7 @@ async def test_bigger_truncation_size(client: openai.AsyncOpenAI):
     expected_message = (
         "truncate_prompt_tokens value is "
         "greater than max_model_len."
-        " Please, select a smaller truncation size."
+        " Please request a smaller truncation size."
     )
     assert error_details["message"] == expected_message
 

tests/entrypoints/pooling/embed/test_online.py

@@ -288,7 +288,7 @@ async def test_truncate_prompt_tokens(client: openai.AsyncOpenAI, model_name: st
         assert "error" in response.object
         assert (
             "truncate_prompt_tokens value is greater than max_model_len. "
-            "Please, select a smaller truncation size." in response.message
+            "Please request a smaller truncation size." in response.message
         )