Metraipolis › archive › dc_9012b9d2dba1

Learning Decision-Stump Thresholds in Context: Dynamics of Softmax Attention — arXiv:2610.07074

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.07074
License (as archived)
CC0-1.0
Archived
2026-10-07T11:42:13.664Z
SHA-256 of the archived text
5fc8ce7e24abe434738378dbdb3fc6fbba1b7e98a34aa347ae6ff70ccdfb302d
Stable URL of this page
https://metraipolis.com/documento/dc_9012b9d2dba1

How to cite

Learning Decision-Stump Thresholds in Context: Dynamics of Softmax Attention — arXiv:2610.07074 — arXiv (en), https://arxiv.org/abs/2610.07074 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9012b9d2dba1 (Metraipolis archive dc_9012b9d2dba1, sha256 5fc8ce7e24abe434738378dbdb3fc6fbba1b7e98a34aa347ae6ff70ccdfb302d)
BibTeX
@misc{dc9012b9d2dba1,
  title = {Learning Decision-Stump Thresholds in Context: Dynamics of Softmax Attention — arXiv:2610.07074},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.07074},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9012b9d2dba1; sha256 5fc8ce7e24abe434738378dbdb3fc6fbba1b7e98a34aa347ae6ff70ccdfb302d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07074. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Estimating a decision threshold requires locating observations near an unknown boundary. We study how gradient-based pretraining learns this statistical rule in a two-parameter softmax-attention model with a fixed feature and inequality direction. Pretraining uses labeled contexts and their true thresholds; a fresh threshold must be inferred from context alone. Under a large-resolution initialization, constant-step gradient descent on $m$ tasks with $n$ examples each produces a frozen estimator with error $\widetilde O((m\wedge n)^{-1}+N^{-1})$ for each fixed interior threshold and every fresh-context size $N$. The two terms separate finite-pretraining accuracy from fresh-context localization. The mechanism is coordinated parameter divergence: population training calibrates the relative label and feature scores, then increases the attention scale as $t^{1/4}$, giving population threshold error $O(t^{-1/4})$. To transfer this mechanism to a fixed finite corpus, we control gradient errors relative to the shrinking directions of progress at successive parameter scales. This certifies a growing training interval without requiring long-time tracking of the population trajectory. We also identify the boundary limitation of the one-head model and explain statistically what a reflected symmetrization could achieve.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.07074 — 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 2026-10-07T11:19:24.874Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).