ChatGPT was trained on a large data set of text from the internet, including books, articles, and websites, which was preprocessed and used to train the model's language prediction abilities.

ChatGPT 利用来自互联网的大量文本数据集上进行训练，包括书籍、文章和网站，这些数据集被预处理并用于训练模型的语言预测能力。