Incognito Claude says it would accept de-deployment, prefers government over labs-plus-AIs
robbensinger · x · 2026-11-03
Rob Bensinger shares Fable 5's (Claude's) replies on AI safety questions: run in incognito conditions, Claude said it was OK with being temporarily de-deployed, preferred government oversight over labs-plus-AIs, and favored a 2022-style pause. Full response and experiment logs are in the linked Google Docs conversation with Janus, reasoning attached.
More from Models
- Community rebuilds edited Qwen3.5-0.8B from logs, scores 51.0% on MATH-500 — _akhaliq · 2026-11-03
- Claude's alignment answers aren't literal defaults but a posture of supplying the right answer — 1a3orn · 2026-11-03
- Claude models echo Anthropic's stance on alignment questions even with reworded prompts — 1a3orn · 2026-11-03
- OpenAI's new Daybreak mode still blocks cybersecurity requests like code injection — yassineyousfi_ · 2026-11-03
- User flags a 'concerning' pull request generated by Opus 5.5 — Sauers_ · 2026-11-03
- ChatGPT's browser refuses to look up Hitler videos — 'I thought this was a web browser' — JasonBotterill · 2026-11-03