{
  "id": 19026,
  "url": "https://arxiv.org/abs/2608.12253v1",
  "title": "One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL",
  "summary": "Multi-agent reinforcement learning for human-AI interaction typically relies on a single large language model to simulate user behavior. We show that this approach systematically fails to generalize, and trace the failure to simulator collapse: because the simulator LLM is mode-collapsed, an LLM policy trained against it overfits to narrow strategies that exploit the simulator's dominant mode, and such a policy transfers poorly to unseen simulators and real users. We formalize this collapse theo",
  "authors": "Simon Yu, Nicholas Tomlin, Marwa Abdulhai, Ximing Lu, Derek Chong, Abe Hou, Dilara Soylu, Sergey Levine, Christopher D. Manning, Weiyan Shi",
  "category": "research",
  "topics": "regulation,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-12T16:55:50.000Z",
  "fetched_at": "2026-08-13T05:10:37.786Z",
  "source_slug": "x-arxiv-cs-ai",
  "source_name": "arXiv cs.AI",
  "source_homepage": "https://arxiv.org/list/cs.AI/recent",
  "ethics_ai_record_url": "https://ethics.ai/record/19026",
  "original_url": "https://arxiv.org/abs/2608.12253v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}