Abstract: The rapid expansion of large language models (LLMs) has led to increasingly frequent interactions between LLM agents and human users, motivating new questions about their capacity to form ...