{
  "id": 5000,
  "url": "https://arxiv.org/abs/2605.03058v2",
  "title": "Neuron-Anchored Rule Extraction for Large Language Models via Contrastive Hierarchical Ablation",
  "summary": "A central goal of explainable AI is to express large language model (LLM) decision logic symbolically and ground it in internal mechanisms. Existing rule-extraction methods usually learn ungrounded symbolic surrogates, while mechanistic interpretability links behavior to neurons but often requires hand-crafted hypotheses and costly interventions. We introduce MechaRule, a pipeline that grounds rule extraction in LLM circuits by localizing sparse agonist activations whose ablation disrupts rule-r",
  "authors": "Francesco Sovrano, Gabriele Dominici, Marc Langheinrich",
  "category": "research",
  "topics": "safety-alignment,transparency",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-04T18:27:37.000Z",
  "fetched_at": "2026-07-14T16:31:26.335Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/5000",
  "original_url": "https://arxiv.org/abs/2605.03058v2",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}