Metraipolis › archive › dc_0d67a9bf04af
From token probabilities to calibrated confidence: An empirical study of mathematical question answering — arXiv:2608.07827
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.07827
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T08:50:39.339Z
- SHA-256 of the archived text
e1d3b0c516070fc5dc5d1b67de67482ac823129ace5a17917b937e7249a9334b- Stable URL of this page
https://metraipolis.com/documento/dc_0d67a9bf04af
How to cite
From token probabilities to calibrated confidence: An empirical study of mathematical question answering — arXiv:2608.07827 — arXiv (en), https://arxiv.org/abs/2608.07827 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0d67a9bf04af (Metraipolis archive dc_0d67a9bf04af, sha256 e1d3b0c516070fc5dc5d1b67de67482ac823129ace5a17917b937e7249a9334b)
BibTeX
@misc{dc0d67a9bf04af,
title = {From token probabilities to calibrated confidence: An empirical study of mathematical question answering — arXiv:2608.07827},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.07827},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0d67a9bf04af; sha256 e1d3b0c516070fc5dc5d1b67de67482ac823129ace5a17917b937e7249a9334b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.07827. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Confidence estimation for large language models (LLMs) aims to estimate the probability that a generated answer is correct, while calibration aligns these estimates with empirical accuracy. Prior work has shown that token probabilities are often overconfident, we investigate whether these readily available signals can nevertheless provide well-calibrated confidence estimation for mathematical question answering. We compare single-pass estimators, which reuse token probabilities from the original generation, with multi-pass estimators, which obtain additional confidence signals through verification or stochastic forward passes. While individual token probabilities can be highly saturated, we find that aggregating token probabilities over the full sequence captures small but consistent differences between correct and incorrect generations, yielding more informative confidence estimates. Multi-pass methods can yield calibrated confidence estimates. We study two such approaches: self-verification through re-prompting, including a lower-cost in-situ variant, and Monte Carlo Dropout, which derives confidence from variation across stochastic forward passes. We further evaluate two post-hoc calibration methods, Platt scaling and isotonic regression, both of which substantially reduce in-domain calibration error. However, their data efficiency varies with dataset difficulty, and the calibration mappings often transfer asymmetrically across datasets and models. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-11 in cs.CL). Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.07827 — 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 2026-10-19T08:49:41.963Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).