{
  "id": 11083797,
  "title": "Semifactual Credit-Augmented Policy Optimization",
  "url": "https://urgent.news/2026/09/30/semifactual-credit-augmented-policy-optimization",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-30T17:59:56.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2609.40360v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "Reinforcement learning with verifiable rewards (RLVR) has improved the reasoning capabilities of large language models (LLMs), yet their predictions remain sensitive to task-irrelevant prompt features. We investigate this sensitivity through semifactual prompt interventions that preserve the underlying problem and its answer. Our analysis reveals substantial variation in token-level sensitivity…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}