What does ChatGPT return about human values? Exploring value bias in ChatGPT using a descriptive value theory