{
  "abstract": "We introduce IrekoGPT, a post-hoc method for converting pretrained LLMs into slimmable models whose width can be adjusted at inference time. Building on SliceGPT, we retain its projection matrices without pruning them, allowing a single model to expose nested subnetworks at different widths. We improve robustness by calibrating each layer across multiple compression ratios, and correct downstream linear layers through gradient-free ridge regression. Across Llama and Qwen models, preliminary results show improvements over naive PCA-based slimming, with the largest gains at high compression. Code is available at https://github.com/aimagelab/IrekoGPT",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2610.00426",
  "author_total": 4,
  "authors": [
    {
      "name": "Pietro Moriello"
    },
    {
      "name": "Pietro Buzzega"
    },
    {
      "name": "Angelo Porrello"
    },
    {
      "name": "Simone Calderara"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/8jkpe0536ejyh.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2610.00426",
  "fulltext_url": null,
  "id": "8jkpe0536ejyh",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2610.00426",
  "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.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1707|Advanced Neural Network Applications"
    ]
  ],
  "openalex": "W7216115615",
  "pdf_url": "https://arxiv.org/pdf/2610.00426",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-01",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/8jkpe0536ejyh.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:26:14Z",
      "native_id": "2610.00426",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2610.00426"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "IrekoGPT: Turning Structured Pruning into Post-Hoc Slimmable LLMs",
  "title_text": "IrekoGPT: Turning Structured Pruning into Post-Hoc Slimmable LLMs",
  "type": "preprint",
  "updated_at": "2026-10-30T12:26:14Z",
  "url": "https://research.draconitic.com/paper/8jkpe0536ejyh",
  "venue": "arXiv",
  "versions": []
}