Metraipolis › archive › dc_6ca6663ab035

System Switch: When Should a Fast Decision Model Stop and Think? — arXiv:2610.09683

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09683
License (as archived)
CC0-1.0
Archived
2026-10-08T11:09:31.736Z
SHA-256 of the archived text
2f9f5035e96191a2adbf78b219ad7a7995717755f49197aa7786f0b7cba05812
Stable URL of this page
https://metraipolis.com/documento/dc_6ca6663ab035

How to cite

System Switch: When Should a Fast Decision Model Stop and Think? — arXiv:2610.09683 — arXiv (en), https://arxiv.org/abs/2610.09683 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6ca6663ab035 (Metraipolis archive dc_6ca6663ab035, sha256 2f9f5035e96191a2adbf78b219ad7a7995717755f49197aa7786f0b7cba05812)
BibTeX
@misc{dc6ca6663ab035,
  title = {System Switch: When Should a Fast Decision Model Stop and Think? — arXiv:2610.09683},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09683},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6ca6663ab035; sha256 2f9f5035e96191a2adbf78b219ad7a7995717755f49197aa7786f0b7cba05812},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09683. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Dual-process agents pair a fast policy with a slow deliberative model. In real-time settings the slow model usually runs continuously; in turn-based agents and robot planners it is invoked on events such as uncertainty or a detected failure. We study a fast learned actor that takes every decision and hands control to a reasoning vision-language model only when a gate opens, while the game keeps running. We use closed-loop Doom and the new open "System One" typed-decision models, served through a common this http URL interface. On 900 held-out questions, (i) zero-shot decision models from 0.15B to 9B parameters choose to collect items 1.6-1.8 times more often than chance among their errors, in any option order, although the order changes some models' accuracy; (ii) accuracy, calibration and sensitivity (how well confidence separates right from wrong answers) are distinct: models of similar accuracy differ widely in AUROC, and the confidence of the most sensitive one tracks which kinds of situation it fails, not which answers are wrong; (iii) offline, deferring the least confident 30% of decisions to a reasoning model gains over random deferral in proportion to the actor's AUROC (rank correlation 0.87); with actor and rate chosen on held-out games the gain is +0.13 [0.08, 0.18] with doomLaya's option order and +0.08 [0.02, 0.14] with shuffled options, and reasoning carries about half of it; (iv) in closed loop (33 games, three seeds) no variant reaches the exit. Committing to plans, the reasoner's or a fixed explore rule's, opens more doors and makes an actor that stands still play; with the rule the agent dies more often. Told that some doors need keys, the reasoner takes ordinary doors for locked ones, which the state cannot tell apart; without that knowledge it goes back to collecting. We release code, prompts, data and logs.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.09683 — 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 2026-10-08T10:14:05.955Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).