{
  "id": 16568,
  "url": "https://arxiv.org/abs/2608.03210v1",
  "title": "ICO: Enhancing Semantic-Shift Jailbreaks via Iterative Context Optimization",
  "summary": "Foundation models have achieved remarkable success across diverse tasks, but they remain vulnerable. To investigate such vulnerabilities, semantic-shift jailbreaks have recently emerged as a promising attack paradigm. They bypass explicit safety mechanisms by replacing harmful terms in original harmful questions with benign alternatives and leveraging contextual information to induce the target model to reinterpret these alternatives as their corresponding harmful concepts. However, existing sem",
  "authors": "Hujian Zhu, Yihao Huang, Felix Juefei-Xu, Xinfeng Li, Peng Zeng, Simeng Qin, Qing Guo, Geguang Pu",
  "category": "research",
  "topics": "safety-alignment,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-04T06:48:17.000Z",
  "fetched_at": "2026-08-05T05:10:44.550Z",
  "source_slug": "x-arxiv-red-teaming-query",
  "source_name": "arXiv red teaming query",
  "source_homepage": "https://arxiv.org/a/redteam",
  "ethics_ai_record_url": "https://ethics.ai/record/16568",
  "original_url": "https://arxiv.org/abs/2608.03210v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}