arxiv preprint - Evaluating Large Language Models as Generative User Simulators for Conversational Recommendation | AI Breakdown | Podwise