Metraipolis › archive › dc_1cc9bc4ebd8b

Aspect-wise Pairwise Tournaments for Test-Time Selection — arXiv:2611.08640

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.08640
License (as archived)
CC0-1.0
Archived
2026-11-09T22:33:38.056Z
SHA-256 of the archived text
6a97f9382f3e34a0a9bbc1876099660d3d0607c7165d7792111532e434eab02a
Stable URL of this page
https://metraipolis.com/documento/dc_1cc9bc4ebd8b

How to cite

Aspect-wise Pairwise Tournaments for Test-Time Selection — arXiv:2611.08640 — arXiv (en), https://arxiv.org/abs/2611.08640 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1cc9bc4ebd8b (Metraipolis archive dc_1cc9bc4ebd8b, sha256 6a97f9382f3e34a0a9bbc1876099660d3d0607c7165d7792111532e434eab02a)
BibTeX
@misc{dc1cc9bc4ebd8b,
  title = {Aspect-wise Pairwise Tournaments for Test-Time Selection — arXiv:2611.08640},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.08640},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1cc9bc4ebd8b; sha256 6a97f9382f3e34a0a9bbc1876099660d3d0607c7165d7792111532e434eab02a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08640. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Scaling test-time compute has proven highly effective at improving LLM reasoning. A popular method is best-of-n sampling, where a model generates many candidate solutions and one is selected. This motivates the question: how can we select the best candidate? When humans make choices, we find it more natural to compare two options than judge each in isolation. Additionally, when the options become complex, we like to judge the options one aspect at a time, such as comparing two job candidates on experience, communication, etc. In the test-time setting, candidate solutions should be judged in the same way. Here we introduce Aspect-wise Pairwise Tournaments (APT), which verifies by comparison and compares by aspect. Each matchup of a single elimination tournament is decided by a vote over multiple off-the-shelf aspect verifiers, each expressing a pairwise preference along its own aspect. Across five domains and three generator LLMs, APT improves over binary multi-aspect approval, self-consistency and reward model verification, with the largest gains on open-ended tasks. We find that both components matter; neither binary approval over the same aspects nor a generic pairwise judge at a matched budget matches APT. Because aspects are specified in natural language, APT also lets users steer selection toward properties they care about, better aligning with user preferences without needing to regenerate solutions. No component is trained, so APT transfers to new domains by simply rewriting prompts.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.08640 — 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 2026-11-09T22:30:47.249Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).