Evaluating the Process Modeling Abilities of Large Language Models -- Preliminary Foundations and Results