david rein
@idavidrein
Leading embedded stress-testing @METR_Evals.
Formerly: early employee @cohere, made GPQA @nyuniversity
Joined July 2018
1.4K Following 5.8K Followers
It seems to me like the Harbor framework for agent evaluations (which is used for Terminal Bench) allows agents to arbitrarily modify their trajectories before they are evaluated. Am I missing something?
Show more
