Evaluating personalized text generated by large language models (LLMs) is challenging, as
only the LLM user, ie, prompt author, can reliably assess the output, but re-engaging the …