Metraipolis › archive › dc_402e06067b2f

When Does the Best Sampling Temperature Rise with the Budget? Sufficient Conditions for Pass@k — arXiv:2608.14665

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.14665
License (as archived)
CC0-1.0
Archived
2026-10-18T19:39:48.730Z
SHA-256 of the archived text
9d2c793e0b0c35364b4f8ba1e3bcd3ae4fefe816cfb9db6dbedf84518c3ab504
Stable URL of this page
https://metraipolis.com/documento/dc_402e06067b2f

How to cite

When Does the Best Sampling Temperature Rise with the Budget? Sufficient Conditions for Pass@k — arXiv:2608.14665 — arXiv (en), https://arxiv.org/abs/2608.14665 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_402e06067b2f (Metraipolis archive dc_402e06067b2f, sha256 9d2c793e0b0c35364b4f8ba1e3bcd3ae4fefe816cfb9db6dbedf84518c3ab504)
BibTeX
@misc{dc402e06067b2f,
  title = {When Does the Best Sampling Temperature Rise with the Budget? Sufficient Conditions for Pass@k — arXiv:2608.14665},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.14665},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_402e06067b2f; sha256 9d2c793e0b0c35364b4f8ba1e3bcd3ae4fefe816cfb9db6dbedf84518c3ab504},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.14665. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The temperature that maximizes pass@$k$ is often low for a small sampling budget and higher for a large budget. This pattern has been reported from Codex through recent multi-sample inference studies. It is not an algebraic property of pass@$k$: as Slocum et al. (ICLR 2025) observe, for one fixed task the maximizing temperature is independent of $k$. Building on that fixed-task observation and the hard/easy-task explanation, we give a formal population-level sufficient condition for the aggregate pattern. For task $X$, let $p_t(X)$ be one-sample success probability at temperature $t$, and define the conditional log-success response $m_t(u)=\mathbb{E}[\dot p_t(X)\mid p_t(X)=u]/u$. If $m_t(u)$ is nonincreasing in current success probability, then the normalized temperature derivative of aggregate pass@$k$ is nondecreasing in $k$. Consequently, derivative signs are nested across budgets; if each temperature-performance curve is strictly single-peaked, its unique maximizer is nondecreasing in $k$. The proof identifies the mechanism as a monotone-likelihood-ratio power tilt toward lower-success tasks. We derive a closed-form two-stratum phase diagram, including upward and downward regimes, and show that the marginal temperature derivative admits an exact $\mathrm{Beta}(2,k)$ kernel representation whose kernel concentrates at one-sample success of order $1/k$. Interpreting that scale as task-level localization additionally requires a regular, nonvanishing density-response factor near zero. A signed-moment representation yields diagnostic shape restrictions, while a short appendix records exact discrete refinements of the existing multi-configuration allocation formulation. No language model is trained, and no model query is used as an experimental measurement: the contribution is a conditional theory of an established empirical phenomenon, with assumptions that can be tested in future work.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.14665 — 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 2026-10-18T19:02:44.038Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).