Automatic Prompt Optimization (ADO) with "Gradient Decent" and Beam Search
Given you have a task $x$, and an Large Language Model $LLM_p(x)$, what is a best prompt $p$ which produces the desired solution $y$?
Given you have a task $x$, and an Large Language Model $LLM_p(x)$, what is a best prompt $p$ which produces the desired solution $y$?