LIMA: Less Is More for Alignment
https://www.lesswrong.com/posts/t9svvNPNmFf5Qa3TA/mysteries-of-mode-collapse
https://crfm.stanford.edu/2023/03/13/alpaca.html
Self-Instruct: Aligning Language Models with Self-Generated Instructions
https://openai.com/blog/chatgpt/