Models That Know How Evaluations Are Designed Score Safer 事件

Models That Know How Evaluations Are Designed Score Safer arXiv:2605.28591v1 Announce Type: new Abstract: The validity of AI safety evaluations depends on models behaving consistently across controlled and deployment settings. Prior work has identified test-time contextual cues, such as hypothetical scenarios, as a source of verbalized evaluation awareness and subsequent behavioral shift. In this paper, we investigate a potential explanation of this phenomenon: evaluation meta-knowledge, define