On  playground.openapi.com   we see a more advanced view of the chat window that chat.openai.com offers (that's the more well-known GUI shown in countlerss screenshots and tutorial videos).

On the Playground GUI, there is a textbox "system" and the default value is "You are a helpful assistant".

Should we enter any personas here in order to make it super clear what we human users intend ChatGPTto do?

(e.g. replace "You are a helpful assistant" with "You are an expert in remote sensing data analysis")

(OpenAI recommends this in their GPT best practices Documentation)

Is this an indication that OpenAI considers this "Persona Pattern" so fundamental that they have added an extra, separate textfield (for specifying the persona) in the structure of the prompt message?

Or is this feature intended just for users of the API , to facilitate the parsing of long-running conversations (even with large intermissions)



You must log in to answer this question.

Browse other questions tagged .