Metraipolis › archive › dc_585679b46a2d

Calibration and Risk-Budgeted Decisions with System One Models — arXiv:2610.19811

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19811
License (as archived)
CC0-1.0
Archived
2026-10-15T14:23:37.690Z
SHA-256 of the archived text
2a6273b4e4a581945fde3c5ea518aea32ca608db0987d5ca42ca527846b8147f
Stable URL of this page
https://metraipolis.com/documento/dc_585679b46a2d

How to cite

Calibration and Risk-Budgeted Decisions with System One Models — arXiv:2610.19811 — arXiv (en), https://arxiv.org/abs/2610.19811 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_585679b46a2d (Metraipolis archive dc_585679b46a2d, sha256 2a6273b4e4a581945fde3c5ea518aea32ca608db0987d5ca42ca527846b8147f)
BibTeX
@misc{dc585679b46a2d,
  title = {Calibration and Risk-Budgeted Decisions with System One Models — arXiv:2610.19811},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19811},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_585679b46a2d; sha256 2a6273b4e4a581945fde3c5ea518aea32ca608db0987d5ca42ca527846b8147f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19811. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLM workflows and agents rely on a set of binary decisions like allow/block an input, or discard/retain a retrieved document. When human deferral is unavailable, uncertain cases must still receive an action. We study System One decision scores in this setting by defaulting to a conservative fallback when uncertain, and controlling the false omission risk within the subset of inputs assigned the utility-maximizing action. Across six datasets covering moderation, safety, and reranking, we compare general System One models with task baselines and a local open-weights fine-tuned System One model. We show that naive baselines for threshold selection exceed the target risk in roughly half of settings with defined test risk, compared with 7.8\% for Learn Then Test (LTT), which leverages finite-sample statistical guarantees. Our results show why relying on a probability-like interface is insufficient for deployment decisions and how risk-budgeted selection trades useful action against observed error.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.19811 — 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 2026-10-15T13:53:35.911Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).