Small Singular Values Matter: A Random Matrix Analysis of Transformer Models

• URL: http://arxiv.org/abs/2410.17770v3
• Date: Thu, 06 Nov 2025 13:25:42 GMT
• Title: Small Singular Values Matter: A Random Matrix Analysis of Transformer Models
• Abstract summary: This work analyzes singular-value spectra of weight matrices in pretrained transformer models.<n>We observe pronounced departures from Random Matrix Theory (RMT) as evidence of randomness and deviations as evidence for learning.<n>Our findings highlight the overlooked importance of the low end of the spectrum and provide theoretical and practical guidance for SVD-based pruning and compression of large language models.
• Score: 5.352699766206808
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This work analyzes singular-value spectra of weight matrices in pretrained transformer models to understand how information is stored at both ends of the spectrum. Using Random Matrix Theory (RMT) as a zero information hypothesis, we associate agreement with RMT as evidence of randomness and deviations as evidence for learning. Surprisingly, we observe pronounced departures from RMT not only among the largest singular values -- the usual outliers -- but also among the smallest ones. A comparison of the associated singular vectors with the eigenvectors of the activation covariance matrices shows that there is considerable overlap wherever RMT is violated. Thus, significant directions in the data are captured by small singular values and their vectors as well as by the large ones. We confirm this empirically: zeroing out the singular values that deviate from RMT raises language-model perplexity far more than removing values from the bulk, and after fine-tuning the smallest decile can be the third most influential part of the spectrum. To explain how vectors linked to small singular values can carry more information than those linked to larger values, we propose a linear random-matrix model. Our findings highlight the overlooked importance of the low end of the spectrum and provide theoretical and practical guidance for SVD-based pruning and compression of large language models.

Related papers

• Making Foundation Models Probabilistic via Singular Value Ensembles [56.4174499669573]
  Foundation models have become a dominant paradigm in machine learning, achieving remarkable performance across diverse tasks through large-scale pretraining.<n>Standard approach to quantifying uncertainty, training an ensemble of independent models, incurs prohibitive computational costs that scale linearly with ensemble size.<n>We propose Singular Value Ensemble (SVE), a parameter-efficient implicit ensemble method that builds on a simple, but powerful core assumption.<n>We show that SVE uncertainty quantification achieves comparable to explicit deep ensembles while increasing the parameter count of the base model by less than 1%.
  arXiv  Detail & Related papers  (2026-01-29T18:07:18Z)
• Efficient Adaptation of Pre-trained Vision Transformer underpinned by Approximately Orthogonal Fine-Tuning Strategy [57.54306942529943]
  We propose an Approximately Orthogonal Fine-Tuning (AOFT) strategy for representing the low-rank weight matrices.<n>Our method achieves competitive performance across a range of downstream image classification tasks.
  arXiv  Detail & Related papers  (2025-07-17T16:09:05Z)
• Spectral Estimation with Free Decompression [47.81955761814048]
  We introduce a novel method of "free decompression" to estimate the spectrum of very large (impalpable) matrices.<n>Our method can be used to extrapolate from the empirical spectral densities of small submatrices to infer the eigenspectrum of extremely large (impalpable) matrices.
  arXiv  Detail & Related papers  (2025-06-13T17:49:25Z)
• Eigenspectrum Analysis of Neural Networks without Aspect Ratio Bias [4.503999875371634]
  Diagnosing deep neural networks (DNNs) through the eigenspectrum of weight matrices has been an active area of research in recent years.<n>We address the impact of the aspect ratio of weight matrices on estimated heavytailness metrics.<n>We propose FARMS, a method that normalizes the weight matrices by subsampling submatrices with a fixed aspect ratio.
  arXiv  Detail & Related papers  (2025-06-06T17:59:28Z)
• Demystifying Singular Defects in Large Language Models [61.98878352956125]
  In large language models (LLMs), the underlying causes of high-norm tokens remain largely unexplored. We provide both theoretical insights and empirical validation across a range of recent models. We showcase two practical applications of these findings: the improvement of quantization schemes and the design of LLM signatures.
  arXiv  Detail & Related papers  (2025-02-10T20:09:16Z)
• How Random Are Ergodic Eigenstates of the Ultrametric Random Matrices and the Quantum Sun Model? [0.0]
  We numerically study the extreme-value statistics of the Schmidt eigenvalues of reduced density matrices obtained from the ergodic eigenstates.<n>For both the ultrametric random matrix and the Quantum Sun model, it can be better described using the extreme value distribution.
  arXiv  Detail & Related papers  (2025-01-31T15:59:18Z)
• Low-Rank Matrix Factorizations with Volume-based Constraints and Regularizations [2.6687460222685226]
  This thesis focuses on volume-based constraints and regularizations designed to enhance interpretability and uniqueness. Motivated by applications such as blind source separation and missing data imputation, this thesis also proposes efficient algorithms.
  arXiv  Detail & Related papers  (2024-12-09T10:58:23Z)
• Reducing the Transformer Architecture to a Minimum [5.352839075466439]
  Transformers are a widespread and successful model architecture, particularly in Natural Language Processing (NLP) and Computer Vision (CV) The attention mechanism itself is nonlinear through its internal use of similarity measures. We have laid the groundwork by testing widespread CV benchmarks: MNIST and CIFAR-10.
  arXiv  Detail & Related papers  (2024-10-17T16:36:14Z)
• 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)
• Transformers are Minimax Optimal Nonparametric In-Context Learners [36.291980654891496]
  In-context learning of large language models has proven to be a surprisingly effective method of learning a new task from only a few demonstrative examples. We develop approximation and generalization error bounds for a transformer composed of a deep neural network and one linear attention layer. We show that sufficiently trained transformers can achieve -- and even improve upon -- the minimax optimal estimation risk in context.
  arXiv  Detail & Related papers  (2024-08-22T08:02:10Z)
• Entrywise error bounds for low-rank approximations of kernel matrices [55.524284152242096]
  We derive entrywise error bounds for low-rank approximations of kernel matrices obtained using the truncated eigen-decomposition. A key technical innovation is a delocalisation result for the eigenvectors of the kernel matrix corresponding to small eigenvalues. We validate our theory with an empirical study of a collection of synthetic and real-world datasets.
  arXiv  Detail & Related papers  (2024-05-23T12:26:25Z)
• Implicit Regularization of Gradient Flow on One-Layer Softmax Attention [10.060496091806694]
  We study gradient flow on the exponential loss for a classification problem with a one-layer softmax attention model. Under a separability assumption on the data, we show that when gradient flow achieves the minimal loss value, it further implicitly minimizes the nuclear norm of the product of the key and query weight matrices.
  arXiv  Detail & Related papers  (2024-03-13T17:02:27Z)
• Weakly supervised covariance matrices alignment through Stiefel matrices estimation for MEG applications [64.20396555814513]
  This paper introduces a novel domain adaptation technique for time series data, called Mixing model Stiefel Adaptation (MSA) We exploit abundant unlabeled data in the target domain to ensure effective prediction by establishing pairwise correspondence with equivalent signal variances between domains. MSA outperforms recent methods in brain-age regression with task variations using magnetoencephalography (MEG) signals from the Cam-CAN dataset.
  arXiv  Detail & Related papers  (2024-01-24T19:04:49Z)
• Unleashing the Power of Pre-trained Language Models for Offline Reinforcement Learning [54.682106515794864]
  offline reinforcement learning (RL) aims to find a near-optimal policy using pre-collected datasets. This paper introduces $textbfLanguage Models for $textbfMo$tion Control ($textbfLaMo$), a general framework based on Decision Transformers to use pre-trained Language Models (LMs) for offline RL. Empirical results indicate $textbfLaMo$ achieves state-of-the-art performance in sparse-reward tasks.
  arXiv  Detail & Related papers  (2023-10-31T16:24:17Z)
• Large-scale gradient-based training of Mixtures of Factor Analyzers [67.21722742907981]
  This article contributes both a theoretical analysis as well as a new method for efficient high-dimensional training by gradient descent. We prove that MFA training and inference/sampling can be performed based on precision matrices, which does not require matrix inversions after training is completed. Besides the theoretical analysis and matrices, we apply MFA to typical image datasets such as SVHN and MNIST, and demonstrate the ability to perform sample generation and outlier detection.
  arXiv  Detail & Related papers  (2023-08-26T06:12:33Z)
• Disentanglement via Latent Quantization [60.37109712033694]
  In this work, we construct an inductive bias towards encoding to and decoding from an organized latent space. We demonstrate the broad applicability of this approach by adding it to both basic data-re (vanilla autoencoder) and latent-reconstructing (InfoGAN) generative models.
  arXiv  Detail & Related papers  (2023-05-28T06:30:29Z)
• Fitting a Directional Microstructure Model to Diffusion-Relaxation MRI Data with Self-Supervised Machine Learning [2.8167227950959206]
  Self-supervised machine learning is emerging as an attractive alternative to supervised learning. In this paper, we demonstrate self-supervised machine learning model fitting for a directional microstructural model. Our approach shows clear improvements in parameter estimation and computational time, compared to standard non-linear least squares fitting.
  arXiv  Detail & Related papers  (2022-10-05T15:51:39Z)
• Random matrix analysis of deep neural network weight matrices [0.0]
  We study the weight matrices of trained deep neural networks using methods from random matrix theory (RMT) We show that the statistics of most of the singular values follow universal RMT predictions. This suggests that they are random and do not contain system specific information.
  arXiv  Detail & Related papers  (2022-03-28T11:22:12Z)
• When Random Tensors meet Random Matrices [50.568841545067144]
  This paper studies asymmetric order-$d$ spiked tensor models with Gaussian noise. We show that the analysis of the considered model boils down to the analysis of an equivalent spiked symmetric textitblock-wise random matrix.
  arXiv  Detail & Related papers  (2021-12-23T04:05:01Z)
• Test Set Sizing Via Random Matrix Theory [91.3755431537592]
  This paper uses techniques from Random Matrix Theory to find the ideal training-testing data split for a simple linear regression. It defines "ideal" as satisfying the integrity metric, i.e. the empirical model error is the actual measurement noise. This paper is the first to solve for the training and test size for any model in a way that is truly optimal.
  arXiv  Detail & Related papers  (2021-12-11T13:18:33Z)
• CNN-based Realized Covariance Matrix Forecasting [0.0]
  We propose an end-to-end trainable model built on the CNN and Conal LSTM (ConvLSTM) It focuses on local structures and correlations and learns a nonlinear mapping that connect the historical realized covariance matrices to the future one. Our empirical studies on synthetic and real-world datasets demonstrate its excellent forecasting ability compared with several advanced volatility models.
  arXiv  Detail & Related papers  (2021-07-22T12:02:24Z)
• Minimax Estimation of Linear Functions of Eigenvectors in the Face of Small Eigen-Gaps [95.62172085878132]
  Eigenvector perturbation analysis plays a vital role in various statistical data science applications. We develop a suite of statistical theory that characterizes the perturbation of arbitrary linear functions of an unknown eigenvector. In order to mitigate a non-negligible bias issue inherent to the natural "plug-in" estimator, we develop de-biased estimators.
  arXiv  Detail & Related papers  (2021-04-07T17:55:10Z)
• SparseBERT: Rethinking the Importance Analysis in Self-attention [107.68072039537311]
  Transformer-based models are popular for natural language processing (NLP) tasks due to its powerful capacity. Attention map visualization of a pre-trained model is one direct method for understanding self-attention mechanism. We propose a Differentiable Attention Mask (DAM) algorithm, which can be also applied in guidance of SparseBERT design.
  arXiv  Detail & Related papers  (2021-02-25T14:13:44Z)
• Sparse Quantized Spectral Clustering [85.77233010209368]
  We exploit tools from random matrix theory to make precise statements about how the eigenspectrum of a matrix changes under such nonlinear transformations. We show that very little change occurs in the informative eigenstructure even under drastic sparsification/quantization.
  arXiv  Detail & Related papers  (2020-10-03T15:58:07Z)
• FLAMBE: Structural Complexity and Representation Learning of Low Rank MDPs [53.710405006523274]
  This work focuses on the representation learning question: how can we learn such features? Under the assumption that the underlying (unknown) dynamics correspond to a low rank transition matrix, we show how the representation learning question is related to a particular non-linear matrix decomposition problem. We develop FLAMBE, which engages in exploration and representation learning for provably efficient RL in low rank transition models.
  arXiv  Detail & Related papers  (2020-06-18T19:11:18Z)
• Efficient MCMC Sampling for Bayesian Matrix Factorization by Breaking Posterior Symmetries [1.3858051019755282]
  We propose a simple modification to the prior choice that provably breaks these symmetries and maintains/improves accuracy. We show that using non-zero linearly independent prior means significantly lowers the autocorrelation of MCMC samples, and can also lead to lower reconstruction errors.
  arXiv  Detail & Related papers  (2020-06-08T00:25:48Z)

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.