I tried my own question using nonsense words – the voodoo incantation worked (without needing the second "and so the answer is..." prompt) @shaneguML @arankomatsuzaki

InstructGPT is 100x smaller (!) than GPT-3 (1.3B params vs 175B) but its outputs are rated higher by human labellers for these tasks. It's somehow nice that the team that kicked off the current GIANT MODELS arms race is also taking steps in a different direction. (3/꩜)

Prompt engineering, one of the most interesting approaches to getting good results from GPT-3, is apparently no longer needed for many tasks with InstructGPT models. Just the question will do. (2/꩜)