Metraipolis › archive › dc_246bcf2f0b36
Dynamic Important Example Mining for Reinforcement Finetuning — arXiv:2608.29252
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.29252
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T07:44:10.546Z
- SHA-256 of the archived text
de65e039b8bbb1da446def57eca87bb17de5424b66cf492136d4a692594d8eb9- Stable URL of this page
https://metraipolis.com/documento/dc_246bcf2f0b36
How to cite
Dynamic Important Example Mining for Reinforcement Finetuning — arXiv:2608.29252 — arXiv (en), https://arxiv.org/abs/2608.29252 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_246bcf2f0b36 (Metraipolis archive dc_246bcf2f0b36, sha256 de65e039b8bbb1da446def57eca87bb17de5424b66cf492136d4a692594d8eb9)
BibTeX
@misc{dc246bcf2f0b36,
title = {Dynamic Important Example Mining for Reinforcement Finetuning — arXiv:2608.29252},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.29252},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_246bcf2f0b36; sha256 de65e039b8bbb1da446def57eca87bb17de5424b66cf492136d4a692594d8eb9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.29252. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement fine-tuning (RFT) is increasingly used to strengthen the reasoning abilities of large models, yet its effectiveness is bound by how training data are selected and used. Most data-centric RFT methods rely on static or heuristic sample selection, implicitly assuming a sample's value is fixed over training. This overlooks the non-stationary dynamics of policy learning and can lead to suboptimal updates. We propose Dynamic Important Example Mining (DIEM), a principled and fully automated framework that makes data utilization adaptive throughout RFT. DIEM integrates two components into each optimization step: (i) a gradient-alignment importance estimator that efficiently approximates each sample's marginal contribution to policy improvement; and (ii) a constrained batch reweighting scheme that maximizes aggregate utility while preserving the update's gradient magnitude to stabilize optimization. Across several reasoning benchmarks, DIEM consistently outperforms strong static and dynamic baselines. The code will be released via this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.29252 — 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 2026-10-17T07:28:26.531Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).