{
  "id": "2611.02333",
  "title": "When Does Adaptive Halting Win? Selection Rules Matter in Looped Language Models",
  "authors": "Mahesh Reddy Pagadala",
  "first_public": "2026-11-02",
  "added_at": "2026-11-03",
  "venue": "arXiv",
  "arxiv": "2611.02333",
  "paper_url": "https://arxiv.org/abs/2611.02333",
  "html_url": "https://arxiv.org/html/2611.02333v1",
  "lens": [
    "recurrent",
    "analysis"
  ],
  "representation": [
    "hidden-state"
  ],
  "computation": [
    "recurrent",
    "depth-wise"
  ],
  "training": [
    "analysis-only",
    "supervised"
  ],
  "questions": [
    "efficiency",
    "evaluation",
    "limitations"
  ],
  "evidence": [
    "empirical",
    "probing"
  ],
  "modality": [
    "language"
  ],
  "summary": "Compares learned and untrained stopping rules on a frozen looped language model. Changing how the stopping threshold is chosen reverses which method saves more loops. Finds no demonstrated advantage for a trained probe over a simple KL-based rule; savings are estimated from stored loop traces.",
  "provenance": "Primary-source literature review, 2026-11-03",
  "author_names": [
    "Mahesh Reddy Pagadala"
  ],
  "record_url": "https://latent-implicit-thinking.github.io/papers/2611.02333/",
  "legacy_record_url": "https://latent-implicit-thinking.github.io/papers/#paper-2611-02333",
  "citation_date": "2026-11-02",
  "json_url": "https://latent-implicit-thinking.github.io/papers/2611.02333/record.json",
  "markdown_url": "https://latent-implicit-thinking.github.io/papers/2611.02333/index.md",
  "bibtex_url": "https://latent-implicit-thinking.github.io/papers/2611.02333/citation.bib",
  "csl_url": "https://latent-implicit-thinking.github.io/papers/2611.02333/citation.csl.json",
  "pdf_url": "https://arxiv.org/pdf/2611.02333"
}