Draft talk:Generative AI and LLM prompt evaluations