1. 如果大规模训练能够带来显著的性能提升，那么伊利亚之前关于“我们正在从规模化配方（预训练）转向规模化思维过程（推理）”的预测可能就错了。 2. 其他实验室复制该模型的成本意味着，目前可能是大多数公众能够接触到最先进模型的最佳时机。从现在开始，成本将呈K曲线式增长，届时普通消费者将使用开源级别的模型。
本文为机器翻译
展示原文
Andrew Curran
@AndrewCurran_
03-29
Three weeks ago there were rumors that one of the labs had completed its largest ever successful training run, and that the model that emerged from it performed far above both internal expectations and what people assumed the scaling laws would predict. At the time these were
来自推特
免责声明：以上内容仅为作者观点，不代表Followin的任何立场，不构成与Followin相关的任何投资建议。
喜欢
收藏
评论
分享
