Transformer Block Coupling and its Correlation with Generalization in LLMs

• URL: http://arxiv.org/abs/2407.07810v3
• Date: Sun, 22 Dec 2024 06:54:57 GMT
• Title: Transformer Block Coupling and its Correlation with Generalization in LLMs
• Authors: Murdock Aubry, Haoming Meng, Anton Sugolov, Vardan Papyan,
• Abstract summary: Large Language Models (LLMs) have made significant strides in natural language processing.<n>We trace the trajectories of individual tokens as they pass through transformer blocks, and linearize the system along these trajectories through their Jacobian matrices.<n>We uncover a $textbftransformer block coupling$ phenomenon in a variety of LLMs, characterized by the coupling of their top singular vectors across tokens and depth.
• Score: 3.007031501305338
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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 trace the trajectories of individual tokens as they pass through transformer blocks, and linearize the system along these trajectories through their Jacobian matrices. By examining the relationships between these Jacobians, we uncover a $\textbf{transformer block coupling}$ phenomenon in a variety 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, namely parameter budget, model depth, and embedding dimension. We further investigate the emergence of these properties through training, noting the development of coupling, as well as an increase in linearity and layer-wise exponential growth in the token trajectories. These collective insights provide a novel perspective on the interactions between token embeddings, and prompt further approaches to study training and generalization in LLMs.

Related papers

• Exact Learning Dynamics of In-Context Learning in Linear Transformers and Its Application to Non-Linear Transformers [1.7034813545878589]
  Transformer models exhibit remarkable in-context learning (ICL) Our work offers an exact dynamical model for ICL and theoretically grounded tools for analyzing complex transformer training.
  arXiv  Detail & Related papers  (2025-04-17T13:05:33Z)
• Sensitivity Meets Sparsity: The Impact of Extremely Sparse Parameter Patterns on Theory-of-Mind of Large Language Models [55.46269953415811]
  We identify ToM-sensitive parameters and show that perturbing as little as 0.001% of these parameters significantly degrades ToM performance. Our results have implications for enhancing model alignment, mitigating biases, and improving AI systems designed for human interaction.
  arXiv  Detail & Related papers  (2025-04-05T17:45:42Z)
• Interpreting Affine Recurrence Learning in GPT-style Transformers [54.01174470722201]
  In-context learning allows GPT-style transformers to generalize during inference without modifying their weights. This paper focuses specifically on their ability to learn and predict affine recurrences as an ICL task. We analyze the model's internal operations using both empirical and theoretical approaches.
  arXiv  Detail & Related papers  (2024-10-22T21:30:01Z)
• Unveiling Induction Heads: Provable Training Dynamics and Feature Learning in Transformers [54.20763128054692]
  We study how a two-attention-layer transformer is trained to perform ICL on $n$-gram Markov chain data. We prove that the gradient flow with respect to a cross-entropy ICL loss converges to a limiting model.
  arXiv  Detail & Related papers  (2024-09-09T18:10:26Z)
• Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models [42.17166746027585]
  We introduce a bidirectional weighted graph-based framework to learn factorized attributes and their interrelations within complex data. Specifically, we propose a $beta$-VAE based module to extract factors as the initial nodes of the graph. By integrating these complementary modules, our model successfully achieves fine-grained, practical and unsupervised disentanglement.
  arXiv  Detail & Related papers  (2024-07-26T15:32:21Z)
• Skip-Layer Attention: Bridging Abstract and Detailed Dependencies in Transformers [56.264673865476986]
  This paper introduces Skip-Layer Attention (SLA) to enhance Transformer models. SLA improves the model's ability to capture dependencies between high-level abstract features and low-level details. Our implementation extends the Transformer's functionality by enabling queries in a given layer to interact with keys and values from both the current layer and one preceding layer.
  arXiv  Detail & Related papers  (2024-06-17T07:24:38Z)
• Dynamical Mean-Field Theory of Self-Attention Neural Networks [0.0]
  Transformer-based models have demonstrated exceptional performance across diverse domains. Little is known about how they operate or what are their expected dynamics. We use methods for the study of asymmetric Hopfield networks in nonequilibrium regimes.
  arXiv  Detail & Related papers  (2024-06-11T13:29:34Z)
• MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities [72.68829963458408]
  We present MergeNet, which learns to bridge the gap of parameter spaces of heterogeneous models. The core mechanism of MergeNet lies in the parameter adapter, which operates by querying the source model's low-rank parameters. MergeNet is learned alongside both models, allowing our framework to dynamically transfer and adapt knowledge relevant to the current stage.
  arXiv  Detail & Related papers  (2024-04-20T08:34:39Z)
• Extending Token Computation for LLM Reasoning [5.801044612920816]
  Large Language Models (LLMs) are pivotal in advancing natural language processing. LLMs often struggle with complex reasoning tasks due to inefficient attention distributions. We introduce a novel method for extending computed tokens in the Chain-of-Thought process, utilizing attention mechanism optimization.
  arXiv  Detail & Related papers  (2024-03-22T03:23:58Z)
• Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimension [63.330262740414646]
  We study how to characterize and predict the truthfulness of texts generated from large language models (LLMs) We suggest investigating internal activations and quantifying LLM's truthfulness using the local intrinsic dimension (LID) of model activations.
  arXiv  Detail & Related papers  (2024-02-28T04:56:21Z)
• Language Models as Hierarchy Encoders [22.03504018330068]
  We introduce a novel approach to re-train transformer encoder-based LMs as Hierarchy Transformer encoders (HiTs) Our method situates the output embedding space of pre-trained LMs within a Poincar'e ball with a curvature that adapts to the embedding dimension. We evaluate HiTs against pre-trained LMs, standard fine-tuned LMs, and several hyperbolic embedding baselines.
  arXiv  Detail & Related papers  (2024-01-21T02:29:12Z)
• CRaSh: Clustering, Removing, and Sharing Enhance Fine-tuning without Full Large Language Model [22.870512676002463]
  This paper focuses on Offsite-Tuning (OFT), a representative technique that transfers transformer blocks between centralized LLMs and downstream emulators. Inspired by these observations, we propose CRaSh, involving Clustering, Removing, and Sharing, a training-free strategy to derive improved emulators from LLMs. Our findings demonstrate a linear connectivity among these optima falling over the same basin, thereby highlighting the effectiveness of CRaSh and OFT.
  arXiv  Detail & Related papers  (2023-10-24T03:08:58Z)
• Seeking Neural Nuggets: Knowledge Transfer in Large Language Models from a Parametric Perspective [106.92016199403042]
  We empirically investigate knowledge transfer from larger to smaller models through a parametric perspective. We employ sensitivity-based techniques to extract and align knowledge-specific parameters between different large language models. Our findings highlight the critical factors contributing to the process of parametric knowledge transfer.
  arXiv  Detail & Related papers  (2023-10-17T17:58:34Z)
• How Do Transformers Learn In-Context Beyond Simple Functions? A Case Study on Learning with Representations [98.7450564309923]
  This paper takes initial steps on understanding in-context learning (ICL) in more complex scenarios, by studying learning with representations. We construct synthetic in-context learning problems with a compositional structure, where the label depends on the input through a possibly complex but fixed representation function. We show theoretically the existence of transformers that approximately implement such algorithms with mild depth and size.
  arXiv  Detail & Related papers  (2023-10-16T17:40:49Z)
• Associative Transformer [25.1364199879332]
  We propose Associative Transformer (AiT) to enhance the association among sparsely attended input tokens. AiT uses a learnable explicit memory comprising specialized priors that guide bottleneck attentions to facilitate the extraction of diverse localized tokens.
  arXiv  Detail & Related papers  (2023-09-22T13:37:10Z)
• Going Beyond Linear Mode Connectivity: The Layerwise Linear Feature Connectivity [62.11981948274508]
  The study of LLFC transcends and advances our understanding of LMC by adopting a feature-learning perspective. We provide comprehensive empirical evidence for LLFC across a wide range of settings, demonstrating that whenever two trained networks satisfy LMC, they also satisfy LLFC in nearly all the layers.
  arXiv  Detail & Related papers  (2023-07-17T07:16:28Z)
• Masked Language Modeling for Proteins via Linearly Scalable Long-Context Transformers [42.93754828584075]
  We present a new Transformer architecture, Performer, based on Fast Attention Via Orthogonal Random features (FAVOR) Our mechanism scales linearly rather than quadratically in the number of tokens in the sequence, is characterized by sub-quadratic space complexity and does not incorporate any sparsity pattern priors. It provides strong theoretical guarantees: unbiased estimation of the attention matrix and uniform convergence.
  arXiv  Detail & Related papers  (2020-06-05T17:09:16Z)
• Multilinear Compressive Learning with Prior Knowledge [106.12874293597754]
  Multilinear Compressive Learning (MCL) framework combines Multilinear Compressive Sensing and Machine Learning into an end-to-end system. Key idea behind MCL is the assumption of the existence of a tensor subspace which can capture the essential features from the signal for the downstream learning task. In this paper, we propose a novel solution to address both of the aforementioned requirements, i.e., How to find those tensor subspaces in which the signals of interest are highly separable?
  arXiv  Detail & Related papers  (2020-02-17T19:06:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.