Metraipolis › archive › dc_341b2f36ff0b

Scaling Parameter and Context in Attention: Native Sparse Attention from Mixture-of-Head — arXiv:2609.38832

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.38832
License (as archived)
CC0-1.0
Archived
2026-10-07T13:13:34.252Z
SHA-256 of the archived text
4ec3eda6de0ee8ef3ef09d06bab20f2e9429dec93eeae48d7ae088eacadf1957
Stable URL of this page
https://metraipolis.com/documento/dc_341b2f36ff0b

How to cite

Scaling Parameter and Context in Attention: Native Sparse Attention from Mixture-of-Head — arXiv:2609.38832 — arXiv (en), https://arxiv.org/abs/2609.38832 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_341b2f36ff0b (Metraipolis archive dc_341b2f36ff0b, sha256 4ec3eda6de0ee8ef3ef09d06bab20f2e9429dec93eeae48d7ae088eacadf1957)
BibTeX
@misc{dc341b2f36ff0b,
  title = {Scaling Parameter and Context in Attention: Native Sparse Attention from Mixture-of-Head — arXiv:2609.38832},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.38832},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_341b2f36ff0b; sha256 4ec3eda6de0ee8ef3ef09d06bab20f2e9429dec93eeae48d7ae088eacadf1957},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38832. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Scaling attention parameters can improve language model quality, but retaining full token histories makes additional heads costly at long contexts. Furthermore, since attention retrieves and combines contextual information, parameter scaling should also support longer contexts. We therefore ask whether attention parameter scaling can directly enable efficient and effective context scaling. We introduce NAMOH, an architecture-native sparse attention mechanism that activates $K$ of $H$ heads per token. Each head retains only its assigned tokens and performs causal attention within this subsequence. Head selection thus jointly determines active parameters and available context without scanning the full history. Under balanced assignments, increasing $H$ at fixed $K$ shortens head histories and reduces per-token key-value (KV) access without increasing total KV storage. We further support head-relative rotary position embeddings to shorten positional spans within routed subsequences, aiming to mitigate position-induced attention noise. Experiments show that NAMOH can outperform fully activated models with the same total parameters, while enabling more efficient long-context inference than smaller dense models with matched active parameter counts. It remains compatible with GQA and existing sparse attention mechanisms. We hope this work offers a new path for scaling attention, with parameter scaling directly enabling context scaling.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.38832 — 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 2026-10-07T12:44:51.770Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).