[BugFix][Frontend] Use LoRA tokenizer in OpenAI APIs (#6227)

Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>

vllm/transformers_utils/tokenizer.py

@@ -88,6 +88,9 @@ def get_tokenizer(
                 "Cannot use the fast tokenizer in slow tokenizer mode.")
         kwargs["use_fast"] = False
 
+    if "truncation_side" not in kwargs:
+        kwargs["truncation_side"] = "left"
+
     try:
         tokenizer = AutoTokenizer.from_pretrained(
             tokenizer_name,