{
  "id": 3862,
  "url": "https://arxiv.org/abs/2605.23652v1",
  "title": "One Policy, Infinite NPCs: Persona-Traceable Shared RL Policies for Scalable Game Agents",
  "summary": "On a 300-persona life-simulation benchmark, pcsp achieves compositional zero-shot persona identification up to 17x above chance, Spearman rho approx 0.73 semantic-behavioral alignment, and 22x faster inference than an LLM-as-policy baseline. Life simulation games require hundreds to thousands of non-player characters (NPCs) that behave consistently with distinct personalities while remaining controllable through designer-authored natural language. Existing methods fail on constraints like person",
  "authors": "Yoosung Hong",
  "category": "research",
  "topics": "regulation,safety-alignment,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-22T14:04:43.000Z",
  "fetched_at": "2026-07-14T16:30:31.924Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/3862",
  "original_url": "https://arxiv.org/abs/2605.23652v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}