natolambert· @natolambert · X·· 5 小时前AI 评分
AAv5指数与开放模型评测之辩
我觉得有必要同时为 AAv5 指数和开放模型辩护。基准评测比大多数人认为的更混乱、更主观，也更棘手。
I feel a need to defend the AAv5 index and open models at the same time. Benchmarking is messier, more opinionated, and trickier than most people give credit to.
来源：natolambert · x.com