HypeDelta
DigestTopicsClaimsPredictionsReliabilityResearchers
Admin
DigestTopicsClaimsPredictionsReliabilityResearchers

HypeDelta - AI Research Intelligence

Claimssafety
safety
fact
neutral

Verbalized eval-awareness in chain-of-thought can be identified as capabilities-flavored or safety-flavored, and these framings predict compliance very differently

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
Artificial Intelligence30 Aug 2026

http://arxiv.org/abs/2608.27340v1