The Stability of Singular Distribution: A Spectral Perspective on the Two-Phase Dynamics of Language Model Pre-training
奇异分布的稳定性:语言模型预训练两阶段动力学的谱视角
机构 * School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences, Beijing, China(中国科学院大学先进交叉学科学院) ; State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China(中国科学院人工智能安全国家重点实验室) ; University of Chinese Academy of Sciences, Beijing, China(中国科学院大学) ; School of Mathematics, Southeast University, Nanjing, China(东南大学数学学院)
专题命中 效率与部署 :language model(title,abstract);large language model(abstract);分类 cs.LG
AI总结 本文发现语言模型预训练中奇异值谱的早期稳定现象(SoSD),并证明该现象与慢下降阶段同步,通过理论分析揭示了权重范数增长导致SoSD阈值,从而限制后续损失下降速率。