Can we get an LLM (GPT-3) to understand or characterize itself by experimenting with itself? aka machine pedagogy. - View it on GitHub