{
  "id": 5642,
  "url": "https://arxiv.org/abs/2604.17691v1",
  "title": "SafeAnchor: Preventing Cumulative Safety Erosion in Continual Domain Adaptation of Large Language Models",
  "summary": "Safety alignment in large language models is remarkably shallow: it is concentrated in the first few output tokens and reversible by fine-tuning on as few as 100 adversarial examples. This fragility becomes critical in real-world deployment, where models undergo sequential adaptation across domains such as medicine, law, and code, causing safety guardrails to erode cumulatively. Yet all existing safety-preserving methods target only single-task fine-tuning, leaving the multi-domain sequential se",
  "authors": "Dongxin Guo, Jikun Wu, Siu Ming Yiu",
  "category": "research",
  "topics": "regulation,safety-alignment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-04-20T01:13:36.000Z",
  "fetched_at": "2026-07-14T16:31:53.168Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/5642",
  "original_url": "https://arxiv.org/abs/2604.17691v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}