AIBearisharXiv – CS AI · 10h ago7/10
🧠
Channel Location Constrains the Auditability of Subliminal Learning
Researchers demonstrate that the auditability of hidden trait transfer in machine learning depends critically on the communication channel through which the trait travels, not merely model size or architecture. Pre-training screens like coverage can detect transfer in initialization-dependent channels but fail against convergent vocabulary geometry in language models, requiring fundamentally different detection approaches.