Refine prompt_template_source: lmms-eval/VLMEvalKit canonical citations

metadata.json

@@ -44,8 +44,9 @@
         }
       },
       "prompt_template_source": {
-        "origin": "fallback",
-        "reference": "T1"
+        "origin": "official",
+        "reference": "https://github.com/EvolvingLMMs-Lab/lmms-eval/blob/main/lmms_eval/tasks/endobench/utils.py (endobench canonical short-answer)",
+        "notes": "lmms-eval EndoBench canonical evaluation prompt."
       }
     }
   }