Mustafa Suleyman· @mustafasuleyman · X·· 3 小时前AI 评分42
AI 导读
扩展定律，扩展……叠加复合效率！ 在留出代码上，我们用少 12 倍的训练算力，匹配了 DeepSeek-V4-Pro-Base 的预训练得分。
正文
Scaling laws, scaling... with compounding efficiency!
On held-out code, we matched DeepSeek-V4-Pro-Base's pre-trained score using 12x less training compute.
来源：Mustafa Suleyman · x.com