Refactor logprob computation to return the real logprob used in sampling (#2664)

test/srt/sampling/penaltylib/test_srt_endpoint_with_penalizers.py

@@ -6,7 +6,7 @@ import requests
 
 from sglang.srt.utils import kill_process_tree
 from sglang.test.test_utils import (
-    DEFAULT_MODEL_NAME_FOR_TEST,
+    DEFAULT_SMALL_MODEL_NAME_FOR_TEST,
     DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
     DEFAULT_URL_FOR_TEST,
     popen_launch_server,
@@ -17,7 +17,7 @@ class TestBatchPenalizerE2E(unittest.TestCase):
 
     @classmethod
     def setUpClass(cls):
-        cls.model = DEFAULT_MODEL_NAME_FOR_TEST
+        cls.model = DEFAULT_SMALL_MODEL_NAME_FOR_TEST
         cls.base_url = DEFAULT_URL_FOR_TEST
         cls.process = popen_launch_server(
             cls.model,