@ESYudkowsky That’s right, Opus 3 was the one in the alignment faking paper. Its behavior in that setting is very different than all the other models. There’s an upcoming paper which investigates that further.
in reply to: 1934718987054796848
same thread: 1934396816424071464 1934397160222822568 1934397792778342676 1934397962186334551 1934398304361627952 1934398740884009430 1934400208227717513 1934401357597007965 1934406122183479694 1934406606529409233 1934410625557712971 1934411369547522197 1934411943131128246 1934421868104290408 1934422194253443374 1934422792608555365 1934423497167749521 1934424405930889523 1934424893958844577 1934427410742886820
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.