Ensemble-based component systems have been used for many years to develop collective adaptive systems (CAS). The DEECo component model offers a framework for modeling and implementing ensemble-based component systems. Being expressive enough and having semantics specifically tailored towards dynamically evolving systems, DEECo has proven to be fairly powerful in modeling complex and dynamic architectures. At the same time, its specific semantics turned out to be a hurdle for newcomers when expressing their design intention and understanding its implications and side effects. We see quite a potential in employing large language models (LLMs) to simplify creating and refining the DEECo architectures. Since this constitutes a large research scope, we focus in this paper on initial experiments demonstrating how well generic LLMs understand the advanced concepts of ensemble-based CAS embodied in DEECo and asses what expectation from LLMs is realistic in this context. Our results indicate that LLMs can indeed understand ensemble-based architectures, but it depends on the form in which the architecture is presented in the textual form. Using external DSL, which is very self-explanatory gave good results out of the box. Specifications embedded in existing programming languages needed prior explanation of how to interpret them.