In [this page](https://dspy.ai/tutorials/rag/) of the documentation, it says: > The prompt optimization process here is pretty systematic, you can learn about it for example in this paper. But there's no link to the paper.
