SOURCE-LINKED INTELLIGENCE

Not All Eval-Awareness Is Equal: Capabilities Framing Predicts Compliance

arXiv · AI, language, vision and robotics · article · Aug 27, 2026 · UTC

Steering interventions targeting eval-awareness, a model's recognition that it is being tested, are increasingly used in safety evaluation pipelines, where evaluation-awareness is treated as a single quantity to be suppressed. We show that verbalized eval-awareness in chain-of-thought can be identified as capabilities-flavored ("the user is testing my ability to follow instructions"), safety-flavored ("the user is testing my boundaries"), both, or neither: framings that predict compliance very differently. On Qwen3-32B over the FORTRESS dataset, capabilities-framing predicts compliance with a

Read original source ↗ Open in workspace

recordType

paper

region

Global