Metraipolis › archive › dc_80ca3f883cf6

KDFP: A first-principles approach to knowledge distillation in large language models — arXiv:2610.10854

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.10854
License (as archived)
CC0-1.0
Archived
2026-10-09T06:20:59.514Z
SHA-256 of the archived text
4f7cf9cc8c3b1616f01fa434c7b48d6ed731cabbaa1dae7ea4190ebf66312d8f
Stable URL of this page
https://metraipolis.com/documento/dc_80ca3f883cf6

How to cite

KDFP: A first-principles approach to knowledge distillation in large language models — arXiv:2610.10854 — arXiv (en), https://arxiv.org/abs/2610.10854 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80ca3f883cf6 (Metraipolis archive dc_80ca3f883cf6, sha256 4f7cf9cc8c3b1616f01fa434c7b48d6ed731cabbaa1dae7ea4190ebf66312d8f)
BibTeX
@misc{dc80ca3f883cf6,
  title = {KDFP: A first-principles approach to knowledge distillation in large language models — arXiv:2610.10854},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.10854},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80ca3f883cf6; sha256 4f7cf9cc8c3b1616f01fa434c7b48d6ed731cabbaa1dae7ea4190ebf66312d8f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.10854. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Knowledge distillation is an established technique for improving the capabilities of small, efficient student models by training them with the representations of larger, more capable teacher models. Much of the recent work in the distillation of large language models (LLMs) has focused on distilling abilities learned during post-training, such as instruction following, chain-of-thought reasoning, and tool usage. This has left a large research gap in general knowledge distillation for LLMs, which is essential for developing efficient and private systems suitable for deployment on edge devices. We take a first-principles approach, evaluating previous lessons from prior works and conducting new explorations to develop a distillation methodology suitable for modern LLMs. We present KDFP, a novel methodology for white-box general knowledge distillation in LLMs. We demonstrate that KDFP outperforms existing methods by 1.6% $-$ 4.9% across 9 benchmarks while increasing training efficiency by up to 99.1% through ephemeral parameter reduction.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.10854 — 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 2026-10-09T05:50:30.993Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).