ChatGPT in particular is trained to predict dialogue and helpful answers to questions, so it would naturally expect some instructions on what exactly it should

ChatGPT in particular is trained to predict dialogue and helpful answers to questions, so it would naturally expect some instructions on what exactly it should do. But for example vanilla GPT-3 has no such bias and merely tries to answer the question "what is the most likely completion to this input based on my corpus?" so it might well simply start generating plausible comments that extend the thread.