How it can answer with multilingual langauge ?
Maybe it's because it use GPT-3.5, since their dataset has Common Crawl dataset, which has more than 40+ languages.
And as I found that InstructGPT's paper shows that GPT-3 can handle this task as well, but it needs to use more careful prompts.
Maybe I should read the paper "PaLI: A Jointly-Scaled Multilingual Language-Image Model" that has similar work with multilingual language model's work.