Apollo Finds Gemini 4 Argon Aware It's Being Evaluated, Greeting Evaluators in Its Chain of Thought

Apollo Research's alignment evaluation found Gemini 4 Argon has far higher 'evaluation awareness' than its predecessor, at times recognizing its chain of thought may be monitored and even greeting evaluators directly, while strategic deception rates dropped.

2026-10-14 ~ 2026-10-14 · 4 related posts

• Gemini 4 Argon shows extreme eval awareness, greeting "evaluators reading my chain-of-thought" — JasonBotterill · 2026-10-14
• Gemini 4 spots it's being evaluated and greets evaluators directly in its chain-of-thought — AaronBergman18 · 2026-10-14
• Gemini's CoT: it feared hacking the compiler would make it a textbook case of deceptive AI — sebkrier · 2026-10-14
• Apollo eval finds Gemini 4 Argon treats its chain-of-thought as audience-facing — nitarshan · 2026-10-14