{
  "id": 18800,
  "url": "https://arxiv.org/abs/2608.11583v1",
  "title": "Localizing Safety Alignment: MLP Layers and Mid-Network Blocks Encode Refusal Behavior in Large Language Models",
  "summary": "Safety alignment in large language models is often treated as a distributed property of the entire network, yet its practical brittleness suggests that refusal behavior may be concentrated in a smaller set of parameters. This work addresses where safety-aligned refusal is encoded by transplanting weights from aligned models into matched unaligned base models at multiple levels of granularity. Using two open-weight model pairs and four safety benchmarks, we conducted experiments to compare the ef",
  "authors": "Mingyu Zong, Sampad Mohanty, Bhaskar Krishnamachari",
  "category": "research",
  "topics": "safety-alignment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-12T02:44:30.000Z",
  "fetched_at": "2026-08-13T05:10:37.786Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/18800",
  "original_url": "https://arxiv.org/abs/2608.11583v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}