Physics of Language Models: Part 3.3,
Knowledge Capacity Scaling Laws
arXiv paper: https://arxiv.org/abs/2404.05405 (last updated April 2024)
Authors: Zeyuan Allen-Zhu and Yuanzhi Li
a short version live-streamed in ICML 2024 tutorial; longer video under plan.
Twitter link for discussions: click here
Slide show (best viewed on a computer)
Slide show (best viewed on a computer)
@article{AL2024-knowledge3,
author = {{Allen-Zhu}, Zeyuan and Li, Yuanzhi},
title = {{Physics of Language Models: Part 3.3, Knowledge Capacity Scaling Laws}},
journal = {ArXiv e-prints},
year = 2024,
month = apr,
volume = {abs/2404.05405},
note = {Full version available at \url{http://arxiv.org/abs/2404.05405}}
}