Large Language Models (LLMs) have made significant strides in natural language processing, and a precise understanding of the internal mechanisms driving their success is essential. In this work, we analyze the trajectories of token embeddings as they pass through transformer blocks, linearizing the system along these trajectories through their Jacobian matrices. By examining the relationships between these block Jacobians, we uncover the phenomenon of \textbf{transformer block coupling} in a multitude of LLMs, characterized by the coupling of their top singular vectors across tokens and depth. Our findings reveal that coupling \textit{positively correlates} with model performance, and that this relationship is stronger than with other hyperparameters such as parameter count, model depth, and embedding dimension. We further investigate how these properties emerge during training, observing a progressive development of coupling, increased linearity, and layer-wise exponential growth in token trajectories. Additionally, experiments with Vision Transformers (ViTs) corroborate the emergence of coupling and its relationship with generalization, reinforcing our findings in LLMs. Collectively, these insights offer a novel perspective on token interactions in transformers, opening new directions for studying their mechanisms as well as improving training and generalization.
Efficient Transformer-based Large Scale Language Representations using Hardware-friendly Block Structured Pruning
University of Connecticut（康涅狄格大学）；Northeastern University（东北大学）；Syracuse University（雪城大学）；Microsoft Corporation（微软公司）；Stevens Institute of Technology（史蒂文斯理工学院）
Use Your INSTINCT: INSTruction optimization for LLMs usIng Neural bandits Coupled with Transformers
National University of Singapore（新加坡国立大学）；Institute of Data Science, National University of Singapore（新加坡国立大学数据科学研究院）；Integrative Sciences and Engineering Programme, National University of Singapore（新加坡国立大学整合科学与工程计划（注：该计划为新加坡国立大学设立的跨学科项目，此处保留机构相关表述））；Tencent（腾讯）；Massachusetts Institute of Technology（麻省理工学院）；Guangdong Lab of AI and Digital Economy (SZ)（广东省人工智能与数字经济实验室（深圳））
Advancing Transformer Architecture in Long-Context Large Language Models: A Comprehensive Survey
State Key Lab of Novel Software Technology, Nanjing University（南京大学计算机软件新技术国家重点实验室）；School of Computing and Mathematical Sciences, Birkbeck, University of London（伦敦大学伯克贝克学院计算与数学科学学院）
Fusing Large Language Models with Temporal Transformers for Time Series Forecasting
University of Science and Technology of China（中国科学技术大学）；University of Washington（华盛顿大学）；Beijing Northern Computility InterConnection Co., Ltd.（北京北方计算互联有限公司）；ENN Group Co., Ltd.（ENN集团有限公司）
ZorBA: Zeroth-order Federated Fine-tuning of LLMs with Heterogeneous Block Activation
Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, Canada（电气与计算机工程系，不列颠哥伦比亚大学）；Department of Computer Science and Engineering, Southern University of Science and Technology, China（计算机科学与工程系，南方科技大学）
Transformer-Assisted LLM-Based Source Code Summarisation: to Enable More Secure Software Development
School of Computing and Communications, Lancaster University, UK（兰卡斯特大学计算与通讯学院）；College of Engineering & Computer Science, VinUniversity（Vin大学工程与计算机科学学院）
The Sharpness Disparity Principle in Transformers for Accelerating Language Model Pre-Training
School of Mathematical Sciences, Peking University（北京大学数学科学学院）；Center for Machine Learning Research, Peking University（北京大学机器学习研究中心）；Shanghai Jiao Tong University（上海交通大学）；AI for Science Institute, Beijing, China（北京人工智能科学研究院）
FlattenGPT: Depth Compression for Transformer with Layer Flattening
State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University（多媒体信息处理国家重点实验室，计算机科学学院，北京大学）；Tsinghua University（清华大学）
Plansformer: Generating Symbolic Plans using Transformers
Artificial Intelligence Institute, University of South Carolina（南卡罗来纳大学人工智能研究所）；IBM Research（IBM研究院）；University of Parma（帕尔马大学）；University of Brescia（布雷西亚大学）
