{
  "abstract": "Large language models (LLMs) as a judge enable scalable evaluation, but their judgments can be sensitive to response order and, even after removing such position effects, can still diverge systematically from human preferences.We introduce DIAL, a unified framework that combines abundant LLM comparisons with limited human comparisons to separate judge-specific position effects, learn shared structure in position-debiased LLM preferences, and adaptively calibrate that structure toward the human preference target. Theoretically, we study three aspects of DIAL: (i) identification of latent LLM preferences, position effects, and human calibration; (ii) adaptive estimation that balances LLM anchoring against limited human evidence; and (iii) fixed-weight uncertainty quantification for the calibrated human preference. Empirically, we evaluate position debiasing and human alignment separately in controlled simulations and on three human-preference benchmarks, showing that DIAL remains robust to unbalanced response order, achieves strong human-aligned rankings with limited labels, and adapts toward human evidence when LLM information is imperfect. Our real-data study collects over 410K judgments from 21 LLM judges in both display orders, providing a resource for future studies of LLM-judge bias, heterogeneity, and human alignment.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.31215",
  "author_total": 4,
  "authors": [
    {
      "name": "Zesheng Cai"
    },
    {
      "name": "Yingqi Fan"
    },
    {
      "name": "Sichang Chen"
    },
    {
      "name": "Jin-Hong Du"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/7wg6hzmmb4fxw.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.31215",
  "fulltext_url": null,
  "id": "7wg6hzmmb4fxw",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.31215",
  "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.AI"
    ],
    [
      "arxiv",
      "stat.AP"
    ],
    [
      "arxiv",
      "stat.ME"
    ],
    [
      "arxiv",
      "stat.ML"
    ]
  ],
  "openalex": null,
  "pdf_url": "https://arxiv.org/pdf/2609.31215",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-25",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/7wg6hzmmb4fxw.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:25:03Z",
      "native_id": "2609.31215",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.31215"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.AP",
      "subfield": "statistics"
    }
  ],
  "title": "DIAL: Position-Debiased LLM Judges with Adaptive Human Preference Calibration",
  "title_text": "DIAL: Position-Debiased LLM Judges with Adaptive Human Preference Calibration",
  "type": "preprint",
  "updated_at": "2026-10-30T12:25:03Z",
  "url": "https://research.draconitic.com/paper/7wg6hzmmb4fxw",
  "venue": "arXiv",
  "versions": []
}