Instruction-following GPT-3 models aligned with reinforcement learning from human feedback.
OpenAI introduced instruction-following models trained with human feedback.
