Metraipolis › archive › dc_9695bca28ca4
Capacity-Dependent Effects of Data Selection for Reasoning — arXiv:2608.13721
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.13721
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T02:21:05.703Z
- SHA-256 of the archived text
3ac38df3626a222f264d73a1ed47e27793cdc48d345fb94ca8777960cad7722e- Stable URL of this page
https://metraipolis.com/documento/dc_9695bca28ca4
How to cite
Capacity-Dependent Effects of Data Selection for Reasoning — arXiv:2608.13721 — arXiv (en), https://arxiv.org/abs/2608.13721 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9695bca28ca4 (Metraipolis archive dc_9695bca28ca4, sha256 3ac38df3626a222f264d73a1ed47e27793cdc48d345fb94ca8777960cad7722e)
BibTeX
@misc{dc9695bca28ca4,
title = {Capacity-Dependent Effects of Data Selection for Reasoning — arXiv:2608.13721},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.13721},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9695bca28ca4; sha256 3ac38df3626a222f264d73a1ed47e27793cdc48d345fb94ca8777960cad7722e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13721. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
In reasoning supervised fine-tuning, candidate responses for the same instruction can differ substantially in how well they match the student's current distribution. Recent likelihood-based response selection methods suggest that responses closer to the student distribution provide more effective supervision, motivating the hypothesis that high-likelihood responses may generally be preferable for fine-tuning. In this paper, we revisit this intuition and show that the value of likelihood-based data selection depends critically on model capacity and training duration. Through controlled experiments on mathematical reasoning, using students ranging from 1.5B to 8B parameters and supervision generated by stronger teacher models, we observe a clear \emph{capacity-dependent} ``{\color{SMALLCOLOR}\textbf{Fast-Fit}} / {\color{LARGECOLOR}\textbf{Slow-Gain}}'' pattern. High-likelihood data provides faster and more stable early improvements, especially for smaller models, but low-likelihood data becomes increasingly beneficial for larger models when training is allowed to continue longer. To explain this phenomenon, we analyze learning dynamics, showing that small models often fail to absorb low-likelihood supervision and instead fall into shallow or repetitive behaviors, while larger models are better able to move toward the teacher distribution under such data. We further provide a capacity-constrained theoretical view of distillation that clarifies how data difficulty, data span, and student capacity jointly govern transfer. Overall, our findings show that effective data selection for reasoning should be aware of model capacity and computing budget rather than based on a single universal preference for high-likelihood supervision.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-17 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.13721 — 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 2026-10-19T02:02:02.862Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).