News for hackers. Stuff that matters.
Re:Sabotage Evaluations for Frontier Models by zb3 on Monday October 21, 2024 @01:37PM (#41906394)
Seems that anthropic can nowadays only compete on "safety", except we don't need it..
Re:Sabotage Evaluations for Frontier Models by Vecr on Monday October 21, 2024 @02:17PM (#41906804)
Sonnet 3.5 is the best model for many tasks.