https://www.landofgpt.com/technique-improves-the-reasoning-capabilities-of-large-language-models/