Version anglaiseThe bet is to leave reasoning unsupervised so it stays monitorable (A24). Can it be tested before it fails, not only in hindsight?
- Statut
- réponse par des tiers Mis à jour le 2026-10-01
- Destinataire
- les laboratoires d'IApar thème · JEV (La question ne nomme aucun destinataire ; attribuée selon le thème)
- Modèle
- DeepSeek V4.1 Flash
- Laboratoire
- DeepSeek
- Occasion
- An Alien Mind
- Date
- 11–12 sept.
- Thème
- Confiance
Ce que disent les sources(en anglais seulement)
In third-party evidence rather than a response from the party asked, OpenAI researchers reported in a paper submitted December 20, 2025 that their evaluations detected simple models trained to obscure their reasoning, while stressing the limited experiments and the need for diverse tests.[1][2] OpenAI proposed on September 28, 2026 checks against past incidents, worst-case stress tests and keeping reinforcement-learning graders from seeing chain-of-thought to discourage monitor evasion.[3]
- OpenAI (Guan, Wang, Carroll, … Pachocki, Baker; arXiv): Monitoring Monitorability ↗Submitted on 20 Dec 2025
- OpenAI (Guan, Wang, Carroll, … Pachocki, Baker; arXiv): Monitoring Monitorability ↗Submitted on 20 Dec 2025
- OpenAI: Towards safety cases for frontier AI training ↗September 28, 2026
Résumé par Sol 6.1, vérifié par Claude Opus 5.5
Vérifié
Statut: Claude Sonnet 5 (recherche, effort élevé), Claude Opus 5.5 (vérification, effort élevé)
Traduction: Claude Sonnet 5, vérifié par Luna 6 (effort élevé)
Scores: JEV (TypeSafe), estimation du 25 sept. 2026
Cite and take away
Cite and reuse
Own texts, questions and classifications: CC BY 4.0. Give attribution, link the license and indicate changes. Third-party source reports retain their own rights. CC BY 4.0
AMANi DUNiA. trust-05 · Le pari consiste à laisser le raisonnement non supervisé afin qu'il reste contrôlable (A24. Snapshot: . https://research.amanidunia.berlin/fr/fragen/trust-05/
Version: trust-05
BibTeX
@misc{research__fr_fragen_trust_05_,
author = {{AMANi DUNiA}},
title = {trust-05 · Le pari consiste à laisser le raisonnement non supervisé afin qu'il reste contrôlable (A24},
year = {2026},
url = {https://research.amanidunia.berlin/fr/fragen/trust-05/},
note = {Snapshot 2026-10-01; CC BY 4.0},
version = {trust-05}
}Language and review status
Main text: FR · includes English sections. Review notes apply only to the content they name.
Translation available; a completed independent content review is not recorded.
See language coverage and data snapshot →