Large language models (LLMs) offer a scalable approach to social simulation, but their credibility depends on how agents are constructed. Existing methods ca…
机构:清华
来源:arXiv 2608.19621 | AI4Papers 论文推荐平台