{
  "abstract": "Self-evolving large language model (LLM) systems repeatedly propose, evaluate, and incorporate updates to prompts, skills, or other persistent artifacts. Despite their growing effectiveness, these systems typically operate under a predetermined iteration or compute budget, without a principled criterion to determine when further evolution is no longer worthwhile. This can lead to two undesirable consequences: unnecessary computation after performance has saturated and the risk of returning late updates that overfit or exploit the evaluation signal. These issues motivate us to study two fundamental questions: when should a self-evolving system stop, and what should it output once it stops? We address the first by formulating an online sequential testing problem and constructing an anytime-valid restart detector using the per-item paired evaluation outcomes already produced by self-evolving LLM systems. We address the second by formulating a change-point estimation problem and using the estimated transition to select an earlier artifact for output. The resulting procedure is plug-and-play and requires no modification of the underlying self-evolving algorithms. Across two self-evolving frameworks, three LLM model families, and five benchmarks, our method substantially reduces computation costs while maintaining comparable unseen-test performance. For example, on SearchQA with SkillOpt and DeepSeek V4 Flash, our method stops at round 4 rather than the full budget of 40, reducing token usage by 91.6% while achieving 82.43% unseen-test accuracy versus 82.00% under the full-budget run.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.04756",
  "author_total": 3,
  "authors": [
    {
      "name": "Enoch Yin"
    },
    {
      "name": "Bin Liu"
    },
    {
      "name": "Zhengling Qi"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/c2943bkm60mr1.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.04756",
  "fulltext_url": null,
  "id": "c2943bkm60mr1",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.04756",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.AI"
    ],
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1702|Topic Modeling"
    ]
  ],
  "openalex": "W7220713001",
  "pdf_url": "https://arxiv.org/pdf/2610.04756",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-03",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/c2943bkm60mr1.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:27:16Z",
      "native_id": "2610.04756",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.04756"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "When Is Enough Enough in Self-Evolving LLM Systems?",
  "title_text": "When Is Enough Enough in Self-Evolving LLM Systems?",
  "type": "preprint",
  "updated_at": "2026-10-30T12:27:16Z",
  "url": "https://research.draconitic.com/paper/c2943bkm60mr1",
  "venue": "arXiv",
  "versions": []
}