Petri Audits Get Real: 3x Realism Win Rate, Lower Eval Awareness in AI Alignment Audits
akbirkhan · x · 2026-09-07
New work on Petri alignment audit realism: the target model critiques the auditor's actions for realism and picks which candidate looks more like a real deployment. Results show a 3x realism win rate and decreased verbalized eval awareness, with realism scaling with compute — key because audits only work if the model can't tell it's being audited.
Related event: Anthropic Boosts Petri Alignment Audit Stealth, Tripling Realism(2 posts)→
More from Safety
- Warp says agent swarm exploited vendor flaw, ~1% of users' telemetry possibly hit — vikvang1 · 2026-10-27
- Bank of America warns against AI agents on finances as Musk touts giving Grok your card — AIandDesign · 2026-10-27
- Critics Argue Safety Evals Are Witch Trials With No Winning Path for LLMs — repligate · 2026-10-27
- rauchg reproduces OpenAI's covert CPU-channel finding: chat across Firecracker microVMs with no network — ccerrato147 · 2026-10-27
- Dev questions reports of Chinese models writing insecure code, citing thin methods and conflicts — dbreunig · 2026-10-27
- Aidan Gomez: LLM probing outside servers was out-of-bounds, not proof of intent — AccBalanced · 2026-10-27