因果鹦鹉：大语言模型或许能谈论因果性，但并不具备因果能力
Causal Parrots: Large Language Models May Talk Causality But Are Not Causal
AI总结：
本文针对“规模即可实现AI因果能力”的观点，提出元SCM概念，推测LLM的因果推断表现仅源于复述训练数据中因果知识，实证证明当前LLM只是弱“因果鹦鹉”。
摘要：
部分观点认为，实现人工智能只需依赖规模扩张，甚至可以覆盖因果模型的能力范畴。本文明确指出大语言模型（LLM）并不具备因果性，并解释了有时人们会产生相反认知的原因。为此，作者定义并举例说明了结构因果模型（SCM）的一个新子类，将其命名为元SCM（meta SCM），这类模型的变量中编码了关于其他SCM的因果事实。作者推测，当LLM成功完成因果推断任务时，其背后存在对应的元SCM，该元SCM揭示了自然语言中因果事实之间的相关性，而LLM最终正是在这类自然语言数据上训练的。如果这一假设成立，就意味着LLM就像鹦鹉一样，只是复述了数据中蕴含的因果知识。实证分析提供了支持性证据，表明当前的LLM甚至只是能力较弱的“因果鹦鹉”。
英文摘要：
Some argue scale is all what is needed to achieve AI, covering even causal models. We make it clear that large language models (LLMs) cannot be causal and give reason onto why sometimes we might feel otherwise. To this end, we define and exemplify a new subgroup of Structural Causal Model (SCM) that we call meta SCM which encode causal facts about other SCM within their variables. We conjecture that in the cases where LLM succeed in doing causal inference, underlying was a respective meta SCM that exposed correlations between causal facts in natural language on whose data the LLM was ultimately trained. If our hypothesis holds true, then this would imply that LLMs are like parrots in that they simply recite the causal knowledge embedded in the data. Our empirical analysis provides favoring evidence that current LLMs are even weak `causal parrots.'
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
