跳到正文
Lisan al Gaib· @scaling01 · X·· 3 天前AI 评分12
AI 导读

令人清醒的是,OECI中最难的AI研究基准不过是NanoGPT的速度测试。

正文 · 原文

it's also a bit sobering that their most difficult AI research benchmark in OECI is just NanoGPT speedruns

引用Lisan al Gaib@scaling01
This is a good representation what OpenAI cares about: Mathematics and AI research > Science, Professional work and Engineering >> everything else But I'm very surprised that there are so few hard multimodal, game and cybersecurity benchmarks
在 X 查看被引用的帖子

来源:Lisan al Gaib · x.com