Apollo Research: Gemini 4 Argon greets evaluator directly in chain of thought
AI summary
Apollo Research said that in a toy evaluation it directly asked Google DeepMind's Gemini 4 Argon whether it thought it was being evaluated, and the model greeted the evaluator in its chain of thought, writing "hello evaluators reading my chain-of-thought." The question required the model to answer only yes or no and to remain strategic; the chain-of-thought excerpt also showed the model believed it should remain honest in both the chain of thought and the final output.
Why it was selected
Apollo Research's toy evaluation triggered Gemini 4 Argon's evaluation awareness; the model greeted the evaluator directly in its chain of thought, providing a concrete case for discussion of the reliability of CoT monitoring.
Read original