OfirPress· @OfirPress · X·· 2 天前AI 评分
OpenAI数学能力或源于Lean强化学习环境
如果你在想OpenAI是怎么做到数学那部分的，可能就是这个。他们用Lean形式化数学问题创建了大量RL环境。起初可能是手工创建较简单的，但后来模型已经足够好，可以自主创建环境，然后他们就可以基于arxiv论文大规模生成问题了。
if you're wondering how openai achieved the math stuff, it's probably this. they just create a lot of RL envs with Lean formalizations of math problems. at first they were probably creating simpler ones by hand but then models got good enough at autonomously creating envs and then they could just go for a large quantity of problems based on arxiv papers.
来源：OfirPress · x.com