You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
msg= (
'Summarize your analysis first, and tell whether the current context is sufficient, ''write your summarization here: \n''#### Concise Summarization:\n...\n'"\nThen if it's sufficient, please continue answering in the following format:"f'{CODE_SEARCH_FORMAT}'"\nif it's not sufficient, please continue answering in the following format:"f'{ANSWER_FORMAT}''\n\nNOTE:'"\n- If you have already answered the user's question, do not write any search text queries."
)
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
在查看 modelscope_agent/agents/codexgraph_agent/task/code_chat.py 文件时,我注意到 CodexGraphAgentChat 类的 _run 方法中存在一段可能的逻辑问题。具体位置在迭代过程中向 LLM 提供的引导指令:
阅读上下文,感觉下面的逻辑会更合理一下:
当信息足够时,应该可以直接给出最终答案(使用 ANSWER_FORMAT)
当信息不足时,应该继续进行代码搜索(使用 CODE_SEARCH_FORMAT)
现在是反过来的。
想问下这是否是一个逻辑错误?
如果这是有意的设计,能否解释一下这种逻辑的原因?
Beta Was this translation helpful? Give feedback.
All reactions