Metraipolis › archive › dc_ddba879c2ade
CORDIAL: Calibrating Ordinal LLM Outputs from Few Labels — arXiv:2609.29807
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.29807
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T04:49:46.563Z
- SHA-256 of the archived text
0e341572ac26d7a1a8ed6ad500589bf5612672e816899ac953618395bb61d2ec- Stable URL of this page
https://metraipolis.com/documento/dc_ddba879c2ade
How to cite
CORDIAL: Calibrating Ordinal LLM Outputs from Few Labels — arXiv:2609.29807 — arXiv (en), https://arxiv.org/abs/2609.29807 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ddba879c2ade (Metraipolis archive dc_ddba879c2ade, sha256 0e341572ac26d7a1a8ed6ad500589bf5612672e816899ac953618395bb61d2ec)
BibTeX
@misc{dcddba879c2ade,
title = {CORDIAL: Calibrating Ordinal LLM Outputs from Few Labels — arXiv:2609.29807},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.29807},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ddba879c2ade; sha256 0e341572ac26d7a1a8ed6ad500589bf5612672e816899ac953618395bb61d2ec},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29807. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A large language model (LLM) can turn a text into a distribution over an ordered scale, but that distribution is a noisy measurement: saturated, compressed or exaggerated, and biased in a consistent direction. We propose CORDIAL, which treats the model's output as a noisy reading of the true label and corrects it with a channel of five interpretable parameters. The channel is small enough for its posterior to be averaged from a handful of labels, and we prove that the resulting calibration preserves first-order stochastic order. On Amazon reviews and CMU-MOSEI transcripts with four LLMs, CORDIAL has the lowest log loss among nine calibrators in 76 of 80 settings with 5 to 100 labels; with 20 labels and the main 7B reader, it matches the strongest baseline using 28-54 labels. The same posterior lets us learn priors from other tasks and fuse several LLMs. Unrestricted calibrators such as Dirichlet calibration overtake it only as the calibration set grows into the hundreds or thousands. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.29807 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-08T04:27:20.104Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).