Least-to-Most Prompting Enables Complex Reasoning in Large Language Models figure
AlphaXiv 中文概览(可滚动查看)