{
  "id": "2610.32492",
  "title": "Accordion: Elastic-Depth Inference via Fixed-Prefix Layer Looping",
  "authors": "Yudong Zhang, Wei Ding, Kaili Zheng, Yuchen Ding, Zhenyu Zhang",
  "first_public": "2026-10-23",
  "added_at": "2026-10-26",
  "venue": "arXiv",
  "arxiv": "2610.32492",
  "paper_url": "https://arxiv.org/abs/2610.32492",
  "html_url": "https://arxiv.org/html/2610.32492v1",
  "lens": [
    "recurrent"
  ],
  "representation": [
    "hidden-state"
  ],
  "computation": [
    "recurrent",
    "depth-wise"
  ],
  "training": [
    "pretraining",
    "supervised",
    "reinforcement-learning",
    "distillation"
  ],
  "questions": [
    "architecture",
    "efficiency",
    "scaling"
  ],
  "evidence": [
    "empirical",
    "scaling"
  ],
  "modality": [
    "language"
  ],
  "summary": "Trains one mixture-of-experts model with five inference budgets by repeating selected layers. Samples a budget during pretraining and preserves all five paths through instruction tuning and reinforcement learning, allowing callers to trade extra computation for stronger predictions without switching models.",
  "provenance": "Primary-source literature review, 2026-10-26",
  "author_names": [
    "Yudong Zhang",
    "Wei Ding",
    "Kaili Zheng",
    "Yuchen Ding",
    "Zhenyu Zhang"
  ],
  "record_url": "https://latent-implicit-thinking.github.io/papers/2610.32492/",
  "legacy_record_url": "https://latent-implicit-thinking.github.io/papers/#paper-2610-32492",
  "citation_date": "2026-10-23",
  "json_url": "https://latent-implicit-thinking.github.io/papers/2610.32492/record.json",
  "markdown_url": "https://latent-implicit-thinking.github.io/papers/2610.32492/index.md",
  "bibtex_url": "https://latent-implicit-thinking.github.io/papers/2610.32492/citation.bib",
  "csl_url": "https://latent-implicit-thinking.github.io/papers/2610.32492/citation.csl.json",
  "pdf_url": "https://arxiv.org/pdf/2610.32492"
}