Metraipolis › archive › dc_04b9d025cddc

Consilience for Verifier-Free Test-Time Scaling — arXiv:2608.09898

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.09898
License (as archived)
CC0-1.0
Archived
2026-10-19T08:50:32.531Z
SHA-256 of the archived text
bd81fe27c16c561e3e5a28300f605fe3bb21179fc5c843759092bed3311435f9
Stable URL of this page
https://metraipolis.com/documento/dc_04b9d025cddc

How to cite

Consilience for Verifier-Free Test-Time Scaling — arXiv:2608.09898 — arXiv (en), https://arxiv.org/abs/2608.09898 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_04b9d025cddc (Metraipolis archive dc_04b9d025cddc, sha256 bd81fe27c16c561e3e5a28300f605fe3bb21179fc5c843759092bed3311435f9)
BibTeX
@misc{dc04b9d025cddc,
  title = {Consilience for Verifier-Free Test-Time Scaling — arXiv:2608.09898},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.09898},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_04b9d025cddc; sha256 bd81fe27c16c561e3e5a28300f605fe3bb21179fc5c843759092bed3311435f9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.09898. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Test-time scaling often uses an external verifier, such as compilers and test cases in coding or trained value functions in robotics applications, to obtain high-quality rollouts. Verifier-free test-time scaling (or VF-TTS) is gaining extensive attention as a mechanism to enhance Large Language Model (LLM) reasoning, primarily because we do not have access to such high-quality verifiers in many real-world applications. Among existing VF-TTS methods, confidence-based VF-TTS methods, which compute and rank rollouts solely by confidence, are particularly promising. Such methods introduce near-zero overhead for sample evaluation and require minimal access to internal model states, making the methods highly flexible across models and tasks. In this paper, we demonstrate a critical limitation of existing confidence-based VF-TTS methods by showing that such methods catastrophically break down on complex tasks. We observe a very interesting phenomenon: uniformly high confidence frequently indicates a failure to explore, favoring confidently wrong answers. To address this, our core insight is that robust cognitive search requires a specific confidence trajectory pattern: such methods perform exploratory branching at the beginning, as manifested by low initial confidence, and converge to a high final confidence solution. To implement this insight, we introduce consilience, a novel selection framework that explicitly evaluates the temporal asymmetry of confidence in reasoning. We operationalize this via a combinatorial metric that actively penalizes high initial confidence while strictly demanding final certainty. Extensive experiments covering both graduate-level mathematics problems and free-form code generation demonstrate that consilience effectively outperforms existing baselines, validating our novel perspective on completion confidence.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-11 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.09898 — 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 2026-10-19T08:49:41.948Z 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).