{
  "id": 7387,
  "url": "https://arxiv.org/abs/2603.10573v1",
  "title": "Implicit Statistical Inference in Transformers: Approximating Likelihood-Ratio Tests In-Context",
  "summary": "In-context learning (ICL) allows Transformers to adapt to novel tasks without weight updates, yet the underlying algorithms remain poorly understood. We adopt a statistical decision-theoretic perspective by investigating simple binary hypothesis testing, where the optimal policy is determined by the likelihood-ratio test. Notably, this setup provides a mathematically rigorous setting for mechanistic interpretability where the target algorithmic ground truth is known. By training Transformers on ",
  "authors": "Faris Chaudhry, Siddhant Gadkari",
  "category": "research",
  "topics": "regulation,safety-alignment,finance-investment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-03-11T09:25:53.000Z",
  "fetched_at": "2026-07-14T16:33:12.390Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/7387",
  "original_url": "https://arxiv.org/abs/2603.10573v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}