{
  "id": 2279128,
  "title": "Phantom Gains: Auditing Self-Improvement Against a Measured Null",
  "url": "https://urgent.news/2026/08/20/phantom-gains-auditing-self-improvement-against-a-measured-null",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-20T17:30:14.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2608.20290v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "Whether a language model has improved itself is increasingly judged not by mean accuracy but by which individual problems it gains and loses. Tracking these transitions means differencing two noisy estimates, leaving them vulnerable to measurement artifacts. Auditing three rounds of rank-$32$ LoRA self-training on Qwen3-8B against a frozen control pushed through the identical pipeline, we…",
  "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."
}