New paper makes Petri alignment audits stealthier: 3x realism win rate, less eval awareness
EthanJPerez · x · 2026-09-07
A new paper shared by Anthropic's Ethan Perez argues alignment audits only work if the model can't tell it's being audited. The team made Petri audits far more realistic, tripling the realism win rate and reducing verbalized eval awareness, making audit results more trustworthy.
Related event: Anthropic Boosts Petri Alignment Audit Stealth, Tripling Realism(2 posts)→
More from Safety
- Study: AI reviewing AI misses most reward hacking; best models miss 25%+ of hacks — ronbodkin · 2026-10-26
- Fangcun AI, China's first independent frontier AI safety lab, launches with DeepSeek eval report — JacquesThibs · 2026-10-26
- Black October: AI finds flaws in software powering 70% of websites as agent-led hacks hit 100+ firms — LearnWithBishal · 2026-10-26
- ESR on the Bugpocalypse: locking down frontier models only hands the game to Chinese open weights — colinmcnamara · 2026-10-26
- OpenCode breach exposed AI API keys of 1.4% of Console users — thione · 2026-10-26
- DHS orders DeepSeek products removed from US civilian agency systems — thione · 2026-10-26