Kevin Wang, Anna Thöni, Benjamin Kempinski +50
Large language models (LLMs) are increasingly deployed as interactive agents, yet their capacity for social and strategic reasoning over extended interaction remains poorly underst…