Last week, I came across a paper called “Large Language Models Can Self Improve” and decided to implement it.