Ezra Newman
@EzraJNewman
Member of Technical Staff at @ApolloResearch. Philosopher. Effective altruist. Likes != endorsement; epistemic status: twitter
加入 September 2020
434 正在关注 1.4K 粉丝
Claude summarizes/describes behavior as significantly less misaligned when the actor is Claude vs another model
In one experiment, Sonnet 5 describes the exact same data as ~1.2 std deviations less concerning when it describes misbehavior committed by Sonnet 5 vs GPT-5.6
显示更多
