Dear GPT community, let me thank you for the opportunity to be here.
I had a chance to run a two-day workshop, and I wanted to be clear about naming. To my surprise, it is quite a challenging problem. And I would like to open a discussion concerning naming specific parts of a text input, which most of us called prompt, but according to the white-paper, a prompt is also a part of the text input.
Text input usually has its structure starts with the task description (sometimes called intent/description/instruction and even a prompt referred in the white-paper in the graph on 4. page)
We have examples that are sometimes called shots, but people coming outside the ML space have hard times understand the term “shot” (my case).
And finally, text the prompt (again, the white-paper referred to this on page 7), a.k.a text that changes, and is a seed for the completion (for a specific type of task as a generation).
What is your opinion? How would you call a text input and its part? When do you say a prompt, what is on your mind?