Metraipolis › archive › dc_83728ca445ae

Copy-Paste Language Models — arXiv:2610.19247

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19247
License (as archived)
CC0-1.0
Archived
2026-10-15T20:02:34.507Z
SHA-256 of the archived text
068e5bc5295cb73d4b920a5d5de2ac110b51e76eba72ca37e8b7caeea9addf5b
Stable URL of this page
https://metraipolis.com/documento/dc_83728ca445ae

How to cite

Copy-Paste Language Models — arXiv:2610.19247 — arXiv (en), https://arxiv.org/abs/2610.19247 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_83728ca445ae (Metraipolis archive dc_83728ca445ae, sha256 068e5bc5295cb73d4b920a5d5de2ac110b51e76eba72ca37e8b7caeea9addf5b)
BibTeX
@misc{dc83728ca445ae,
  title = {Copy-Paste Language Models — arXiv:2610.19247},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19247},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_83728ca445ae; sha256 068e5bc5295cb73d4b920a5d5de2ac110b51e76eba72ca37e8b7caeea9addf5b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19247. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Transformers are intuitively considered to be efficient at copying from their inputs. We actually show that the ability of LLMs to copy in context can be improved through explicit modeling. We propose Copy-Paste Language Models (CPLMs), which add a lightweight Copy-Paste head to the conventional language modeling (LM) head. CPLMs improve training efficiency, downstream performance, and offer new operating points in long-context and retrieval setups. We validate our approach by pretraining models from 135M to 1.7B parameters, showing improvements across all scales and backbones with negligible training and inference overheads. We explain the success of CPLMs by showing that they partially mitigate the LM head gradient bottleneck through an alternative gradient path.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.19247 — 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 2026-10-15T19:35:48.324Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).