{
  "id": 12988,
  "url": "https://arxiv.org/abs/2607.17890v2",
  "title": "Stress Testing Concept Erasure with Large Language Model Agents",
  "summary": "Concept erasure aims to remove semantic concepts from a trained generative model and is increasingly important for responsible AI deployment. However, verifying whether a model has robustly removed targeted concepts remains a critical challenge. Existing evaluation methods are typically pre-defined and static, failing to expose vulnerabilities under diverse natural-language probes and challenging conditions. Moreover, manually designed evaluation strategies can be biased and difficult to scale.",
  "authors": "Yuyang Xue, Feng Chen, Zhihua Liu, Edward Moroshko, Jingyu Sun, Steven McDonagh, Sotirios A. Tsaftaris",
  "category": "research",
  "topics": "bias-fairness,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-07-20T12:38:50.000Z",
  "fetched_at": "2026-07-23T05:10:49.458Z",
  "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/12988",
  "original_url": "https://arxiv.org/abs/2607.17890v2",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}