better：进一步拆分模型配置

src/experimental/PFC/pfc.py

@@ -44,7 +44,7 @@ class GoalAnalyzer:
     def __init__(self, stream_id: str, private_name: str):
         # TODO: API-Adapter修改标记
         self.llm = LLMRequest(
-            model=global_config.model.normal, temperature=0.7, max_tokens=1000, request_type="conversation_goal"
+            model=global_config.model.utils, temperature=0.7, max_tokens=1000, request_type="conversation_goal"
         )
 
         self.personality_info = individuality.get_prompt(x_person=2, level=3)