{
  "abstract": "The repetition count that works best for a small language model may not remain best at a larger scale. We study this effect in pretraining with a finite target corpus mixed with generic data at a fixed target fraction. On Wikipedia-derived data and Proof-Pile-2, the ranking of measured repetition counts changes with model size, and a 520M Proof-Pile-2 experiment confirms that reducing repetition from sixteen to eight improves loss while using fewer training tokens. We use loss curves from several smaller models to retain a short list of promising repetition counts for evaluation at a larger scale. On PubMed and Caselaw, candidate sets fixed before target-model training retain the lowest-loss measured count on the original evaluation grids at both 200M and 520M. This supports candidate retention as a practical alternative to exact point prediction. We also relate the pruning regression to an empirical scaling model with two opposing repetition-dependent loss terms. A first-order expansion in log model size yields the linear form used by the selection rule, providing a scaling-based interpretation of the candidate-selection procedure.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.05126",
  "author_total": 2,
  "authors": [
    {
      "name": "Ziyue WANG"
    },
    {
      "name": "T. Kanamori"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/84eeppcx79ngs.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.05126",
  "fulltext_url": null,
  "id": "84eeppcx79ngs",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.05126",
  "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.CL"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1702|Topic Modeling"
    ]
  ],
  "openalex": "W7220732105",
  "pdf_url": "https://arxiv.org/pdf/2610.05126",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-04",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/84eeppcx79ngs.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:27:16Z",
      "native_id": "2610.05126",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.05126"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Selecting Repetition Counts Across Model Scales in Data-Constrained Pretraining",
  "title_text": "Selecting Repetition Counts Across Model Scales in Data-Constrained Pretraining",
  "type": "preprint",
  "updated_at": "2026-10-30T12:27:16Z",
  "url": "https://research.draconitic.com/paper/84eeppcx79ngs",
  "venue": "arXiv",
  "versions": []
}