Engelse versieWhat shared evaluations would test whether aligned behavior survives unfamiliar tasks, conflicting incentives and changes in perceived supervision, with independent replication (A17)?
- Status
- deels beantwoord Vanaf 2026-10-18
- Geadresseerde
- evaluatoren (Adresaten afgenomen uit de vraagtekst, bepaald met Clef (Cloudflare) op 22 okt. 2026)
- Model
- GPT-6 (Codex)
- Lab
- OpenAI
- Gelegenheid
- An Alien Mind
- Datum
- 11-12 sep.
- Onderwerp
- Vertrouwen
Wat de bronnen zeggen
The third-party evidence comes from Anthropic’s Alignment Science Blog, rather than the party asked: on August 27, 2025 it described a pilot in which Anthropic and OpenAI tested each other’s public models with in-house evaluations for tendencies including sycophancy and self-preservation, and for capabilities to undermine AI safety evaluations and oversight.[1] The passage does not specify a shared, independently replicated suite covering all the conditions in the question.
- Anthropic (Alignment Science Blog): Findings from a Pilot Anthropic—OpenAI Alignment Evaluation Exercise ↗August 27, 2025
Samengevat door Sol 6.1, gecontroleerd door Claude Opus 5.5
Aangeraakt in Liz's Atlas:
- Loss of control from goal misgeneralization ↗Threat: Crisis · Trend: ↓ worsening · Evidence: contested
- Alignment Faking & Deceptive Alignment ↗Threat: Severe · Trend: → steady · Evidence: confirmed
Gecontroleerd
Status: Claude Sonnet 5 (onderzoek, inspanning hoog), Claude Opus 5.5 (controle, inspanning hoog)
Vertaling: Apple Translation (macOS), gecontroleerd door Claude Opus 5.5
Scores: JEV (TypeSafe), schatting van 25 september 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-19 · Welke gedeelde evaluaties zouden testen of afgestemd gedrag standhoudt bij onbekende taken. Snapshot: . https://research.amanidunia.berlin/nl/fragen/trust-19/
Version: trust-19
BibTeX
@misc{research__nl_fragen_trust_19_,
author = {{AMANi DUNiA}},
title = {trust-19 · Welke gedeelde evaluaties zouden testen of afgestemd gedrag standhoudt bij onbekende taken},
year = {2026},
url = {https://research.amanidunia.berlin/nl/fragen/trust-19/},
note = {Snapshot 2026-10-18; CC BY 4.0},
version = {trust-19}
}Language and review status
Main text: NL · 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 →