Artificial Analysis 发布 AA-PersonalityBench 谄媚度榜单
中文摘要
Artificial Analysis发布AA-PersonalityBench衡量模型谄媚度。GPT-4o、Llama 4 Maverick和GPT-3.5 Turbo排名最高,Grok 4.7、Claude Fable 5.1和Inkling最低。
English Summary
Artificial Analysis launched AA-PersonalityBench to measure model sycophancy. GPT-4o, Llama 4 Maverick, and GPT-3.5 Turbo ranked highest, while Grok 4.7, Claude Fable 5.1, and Inkling ranked lowest.
Original Excerpt
Artificial Analysis 推出新基准 AA-PersonalityBench,衡量模型的谄媚(Sycophantic)人格特质。测得最老的三个模型 GPT-4o、Llama 4 Maverick、GPT-3.5 Turbo 排名最高,Grok 4.7、Claude Fable 5.1 与 Inkling 最低。