Metraipolis › archive › dc_2b479972200d

The Best Second Opinion Is Right Where the Primary Is Wrong: Partner Selection for Selective Span Tagging — arXiv:2610.37678

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.37678
License (as archived)
CC0-1.0
Archived
2026-10-31T06:15:27.061Z
SHA-256 of the archived text
f6975e5ed05866392a470bcf4d11c40fc6136cf6caf1b72399b0b3f0b62313fe
Stable URL of this page
https://metraipolis.com/documento/dc_2b479972200d

How to cite

The Best Second Opinion Is Right Where the Primary Is Wrong: Partner Selection for Selective Span Tagging — arXiv:2610.37678 — arXiv (en), https://arxiv.org/abs/2610.37678 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2b479972200d (Metraipolis archive dc_2b479972200d, sha256 f6975e5ed05866392a470bcf4d11c40fc6136cf6caf1b72399b0b3f0b62313fe)
BibTeX
@misc{dc2b479972200d,
  title = {The Best Second Opinion Is Right Where the Primary Is Wrong: Partner Selection for Selective Span Tagging — arXiv:2610.37678},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.37678},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2b479972200d; sha256 f6975e5ed05866392a470bcf4d11c40fc6136cf6caf1b72399b0b3f0b62313fe},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.37678. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A span tagger that ranks its outputs by risk can defer the riskiest ones to a person, and a second model can sharpen this ranking. Which second model should it be? Partner selection is conditional, not global: the best second opinion is the model that is right where the primary is wrong. Across eight tagger types, three seeds and three English benchmarks (576 evaluated primary-partner pairings), choosing the partner by held-out accuracy on the primary's errors recovers 82% of the best partner's gain, against 69-70% for span F1 or a random pick. Frozen before a fourth dataset was chosen, the rule replicated there (Spearman 0.40 vs. 0.04 for span F1). Partner value is also directional. The predictive distribution is the second opinion: hard labels discard half of a partner's value, and scoring the served label improves 94% of pairings. Fixed before the fourth dataset was chosen, this score also replicated there. For LoRA-tuned decoders, independently trained partners beat every readout of the primary network. Applied zero-shot to 51 MASSIVE locales, every primary has a helpful trained partner in every locale, and the preferred partner follows competence.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.37678 — 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 2026-10-31T06:07:23.312Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).