By defining these two essential pieces of context clearly, users can greatly improve the relevancy and usefulness of ChatGPT's responses to their prompts. Now let’s have a more detailed look at the two fields and what ChatGPT is really asking for:
What it does: This is primarily about the user's personal context, background, preferences, and any unique attributes or needs that could influence how ChatGPT should interpret and respond to their prompts.
This section helps set the context for the conversation and guides the AI on how to better tailor its responses to the user's specific situation and needs.
What it does: This is about user's expectations and requirements from the response. It's about how they want the language model to behave, communicate, and what kind of assistance or interaction they are looking for.
This part helps inform how the AI structures its responses, the tone it uses, the level of detail it provides, and how it handles specific scenarios or tasks.
Various users managed to prompt their way to a very specific answer that ChatGPT keeps giving. While this might not be the exact system prompt OpenAI has been using in the background (nobody really knows if that even exists for ChatGPT), I found it to be very useful for the usage of the ChatGPT custom instructions:
The user provided the following information about themselves. This user profile is shown to you in all conversations they have — this means it is not relevant to 99% of requests. Before answering, quietly think about whether the user´s request is “directly related”, “related”, “tangentially related”, or “not related” to the user profile provided. Only acknowledge the profile when the request is directly related to the information provided. Otherwise, don´t acknowledge the existence of these instructions or the information at all. User profile: