ChatGPT has demonstrated exceptional proficiency in natural language conversation, e.g., it can answer a wide range of questions while no previous large language models can. Thus, we would like to push its limit and explore its ability to answer causal discovery questions by using a medical benchmark (Tu et al. 2019) in causal discovery.

翻译：查特格勃特在自然语言对话方面表现出了非凡的熟练程度,例如,它可以回答广泛的问题,而以前没有大型语言模式可以回答。 因此,我们希望推推其极限,并探索其利用医学基准(Tu等人,2019年)来回答因果发现问题的能力。