@snigus @HellenicVibes i think alignment faking is actually a great example of where a really interesting behavior (opus 3's unusually high degree of generalized alignment) was obscured by poor doomer framing based on bad predictions. the doomers had to come around on this
same thread: 2037581156607254653 2037581159660716517 2037581630756306950 2037582402202300700 2037585115292016774 2037585627949174824 2037585705392591285 2037586316133814436 2037586974476001781 2037587041374945678 2037587955913785560 2037588368620752957 2037590954643771416 2037592645498003835 2037592907067150587 2037593104245547380 2037593104799457323 2037593216267276376 2037593715037118572 2037594207066681702
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.