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：ChatGPT在自然语言对话中展现出卓越的能力，例如，它能回答广泛的问题，而此前的任何大语言模型都无法做到。因此，我们希望突破其极限，通过使用因果发现领域的医学基准（Tu 等人，2019），探索其在回答因果发现问题方面的能力。