1 paper
Hanwen Xing, Pengyun Wang, BingXu Meng +8
Progress in large language models is often summarized using a single scalar measure, such as a time horizon, a latent ability estimate, or an aggregate benchmark score. These summa…