BriefGPT.xyz
May, 2024
通过从头开始训练领域知识来匹配领域专家
Matching domain experts by training from scratch on domain knowledge
HTML
PDF
Xiaoliang Luo, Guangzhi Sun, Bradley C. Love
TL;DR
通过领域特定的自回归训练方法,在小型的GPT-2模型上训练神经科学文献,可以实现即使在小规模的语言模型上也能达到专家水平的性能表现。
Abstract
Recently,
large language models
(LLMs) have outperformed human experts in predicting the results of
neuroscience experiments
(Luo et al., 2024). What is the basis for this performance? One possibility is that
→