Metraipolis › archive › dc_cbe4c19e2fc6

LEV: Thinking Fast and Deep with Looped Decision Models — arXiv:2610.20951

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.20951
License (as archived): CC0-1.0
Archived: 2026-10-16T08:59:34.102Z
SHA-256 of the archived text: 0ed994ab94ea395d63139ebecf6020c4fa0cf3d784c2bdc5af7840801441cd77
Stable URL of this page: https://metraipolis.com/documento/dc_cbe4c19e2fc6

How to cite

LEV: Thinking Fast and Deep with Looped Decision Models — arXiv:2610.20951 — arXiv (en), https://arxiv.org/abs/2610.20951 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cbe4c19e2fc6 (Metraipolis archive dc_cbe4c19e2fc6, sha256 0ed994ab94ea395d63139ebecf6020c4fa0cf3d784c2bdc5af7840801441cd77)

BibTeX

@misc{dccbe4c19e2fc6,
  title = {LEV: Thinking Fast and Deep with Looped Decision Models — arXiv:2610.20951},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20951},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cbe4c19e2fc6; sha256 0ed994ab94ea395d63139ebecf6020c4fa0cf3d784c2bdc5af7840801441cd77},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20951. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Decision models answer typed questions about an input with one probability per option, computed in a single forward pass without generating text. Like the fast System 1 of dual-process theory, they answer at once, but their depth is fixed at training time, so they spend the same computation on easy and hard questions. Models that reason in generated text can think longer, but they pay for every extra step in decoded tokens. We present LEV, a decision model that thinks deeper by rerunning its layers in a loop instead of generating tokens. LEV builds on a language model pretrained as a loop, reads the options after every loop step and trains all steps jointly, so one checkpoint can answer at any depth. On three looped backbones, accuracy rises with depth on the same weights, mostly on hard questions, and a 2.6B LEV goes from 63.2% at depth one to 76.2% at depth four on JevBench. On-demand depth, set per question by a small stopper with a threshold fitted on unlabelled requests, saves 15% to 24% of the computation at a cost of under one accuracy point, level with or slightly above a fixed mix of depths. On a locked test set, the 1.4B and 3B models are 5.4 to 8.8 points more accurate than dense (non-looped) models and cascades of the same latency below 50 ms.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.20951 — 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 2026-10-16T08:44:17.679Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).