{
  "id": "2505.18642",
  "title": "Skip-Thinking: Chunk-wise Chain-of-Thought Distillation Enable Smaller Language Models to Reason Better and Faster",
  "authors": "Xiaoshu Chen, Sihang Zhou, Ke Liang, Xiaoyu Sun, Xinwang Liu",
  "first_public": "2025-05-24",
  "added_at": "2026-11-02",
  "venue": "arXiv",
  "arxiv": "2505.18642",
  "paper_url": "https://arxiv.org/abs/2505.18642",
  "html_url": "https://arxiv.org/html/2505.18642v1",
  "lens": [
    "internalized"
  ],
  "representation": [
    "text-latent-hybrid"
  ],
  "computation": [
    "token-wise",
    "depth-wise"
  ],
  "training": [
    "distillation",
    "supervised"
  ],
  "questions": [
    "capability",
    "efficiency"
  ],
  "evidence": [
    "empirical"
  ],
  "modality": [
    "language"
  ],
  "summary": "Splits teacher rationales into chunks, then trains a student to skip chunks whose removal preserves the answer. The model retains key written steps while internalizing transitional or summary chunks.",
  "provenance": "Primary-source literature review, 2026-11-02",
  "author_names": [
    "Xiaoshu Chen",
    "Sihang Zhou",
    "Ke Liang",
    "Xiaoyu Sun",
    "Xinwang Liu"
  ],
  "record_url": "https://latent-implicit-thinking.github.io/papers/2505.18642/",
  "legacy_record_url": "https://latent-implicit-thinking.github.io/papers/#paper-2505-18642",
  "citation_date": "2025-05-24",
  "json_url": "https://latent-implicit-thinking.github.io/papers/2505.18642/record.json",
  "markdown_url": "https://latent-implicit-thinking.github.io/papers/2505.18642/index.md",
  "bibtex_url": "https://latent-implicit-thinking.github.io/papers/2505.18642/citation.bib",
  "csl_url": "https://latent-implicit-thinking.github.io/papers/2505.18642/citation.csl.json",
  "pdf_url": "https://arxiv.org/pdf/2505.18642"
}