Metraipolis › archive › dc_c19dec3ee826

Strategic Self-Consistency — arXiv:2609.30352

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.30352
License (as archived): CC0-1.0
Archived: 2026-10-08T01:38:51.271Z
SHA-256 of the archived text: c8b99b39e891252e755105aacf814ca87fde760352ebb61375fad58be9edd886
Stable URL of this page: https://metraipolis.com/documento/dc_c19dec3ee826

How to cite

Strategic Self-Consistency — arXiv:2609.30352 — arXiv (en), https://arxiv.org/abs/2609.30352 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c19dec3ee826 (Metraipolis archive dc_c19dec3ee826, sha256 c8b99b39e891252e755105aacf814ca87fde760352ebb61375fad58be9edd886)

BibTeX

@misc{dcc19dec3ee826,
  title = {Strategic Self-Consistency — arXiv:2609.30352},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.30352},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c19dec3ee826; sha256 c8b99b39e891252e755105aacf814ca87fde760352ebb61375fad58be9edd886},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30352. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Self-consistency has become a popular technique for enhancing the reasoning abilities of large language models by generating multiple reasoning paths and selecting the final answer through a majority vote. However, because model providers typically charge users in proportion to the number of reasoning paths generated, they have a financial incentive to artificially increase the path count. In this work, we show that an unfaithful provider can exploit this incentive using a simple, efficient algorithm while avoiding detection by an auditor: by generating and strategically reordering additional reasoning paths, the algorithm makes every path appear necessary to reach the majority. To validate our algorithm, we conduct experiments with multiple instruct models from the Llama and Qwen families, as well as reasoning models distilled from DeepSeek-R1, on benchmark datasets spanning mathematics, science, and question answering. Our results suggest that the distribution of additional reasoning paths generated by our algorithm is heavy-tailed and that substantial capacity to overcharge remains even under the best possible audit designed to keep the false-positive rate below $\alpha = 0.1$.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2609.30352 — 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 2026-10-08T01:12:14.419Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).