Metraipolis › archive › dc_77d0abffedfc

POOL: Propagated Uncertainty Over Lookalikes — arXiv:2608.23086

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.23086
License (as archived): CC0-1.0
Archived: 2026-10-18T07:41:50.631Z
SHA-256 of the archived text: 599b05412961ee3da011ec13290787a017006e5eae27cacd259f27fc84396925
Stable URL of this page: https://metraipolis.com/documento/dc_77d0abffedfc

How to cite

POOL: Propagated Uncertainty Over Lookalikes — arXiv:2608.23086 — arXiv (en), https://arxiv.org/abs/2608.23086 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_77d0abffedfc (Metraipolis archive dc_77d0abffedfc, sha256 599b05412961ee3da011ec13290787a017006e5eae27cacd259f27fc84396925)

BibTeX

@misc{dc77d0abffedfc,
  title = {POOL: Propagated Uncertainty Over Lookalikes — arXiv:2608.23086},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.23086},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_77d0abffedfc; sha256 599b05412961ee3da011ec13290787a017006e5eae27cacd259f27fc84396925},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.23086. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Black-box large language models need confidence scores that can separate likely-correct from likely-incorrect outputs, enabling systems to prioritize human review, route uncertain cases to stronger models, or choose abstention thresholds on development data. Yet existing confidence estimators face a cost-quality trade-off: verbal confidence is cheap but is often overconfident, while sampling-based uncertainty is more informative but scales linearly with the number of samples per query. We propose \textsc{POOL} (\emph{Propagated Uncertainty Over Lookalikes}),a cost-efficient framework that addresses this trade-off taking inspiration from group-testing.\textsc{POOL} clusters query stems with overlaps, evaluates a base estimator on representative medoids, softly propagates confidence scores to nearby queries, and selectively evaluates high-disagreement cases. We instantiate this framework with \textsc{Hy@}$p$, a hybrid estimator that combines verbal confidence with spectral answer diversity computed from the negative von Neumann entropy of sampled answer this http URL six domains from three datasets and five black-box LLMs, \textsc{Hy@}5 achieves higher average AUROC than verbal confidence and \textsc{Vn@}10 sampling while using half as many samples as \textsc{Vn@}10. \textsc{POOL}-\textsc{Hy@}5 retains 93.5--97.9\% of its AUROC while saving 19.3--39.3\% of generations. On paraphrase-dense workloads, generation savings rise to 73-76\%, showing that semantic redundancy can be leveraged to lower confidence-estimation costs.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.23086 — 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 2026-10-18T07:37:52.767Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).