Update tokenizer_config.json

tokenizer_config.json

@@ -118,6 +118,6 @@
     "attention_mask"
   ],
   "model_max_length": 2048,
-  "pad_token": "<|endoftext|>",
+  "pad_token": ">>SUFFIX<<",
   "tokenizer_class": "PreTrainedTokenizerFast"
 }