Has anyone compared DeepSeek on mixed Chinese-English prompts, especially technical questions where the explanation is in one language but package names, error messages, or code comments are in the other?

I’m curious whether it tends to keep the response in the language of the main question, and whether terminology or nuance gets less reliable when both languages are used in the same prompt.