S2D: Selective Spectral Decay for Quantization-Friendly Conditioning of Neural Activations

• URL: http://arxiv.org/abs/2602.14432v1
• Date: Mon, 16 Feb 2026 03:41:06 GMT
• Title: S2D: Selective Spectral Decay for Quantization-Friendly Conditioning of Neural Activations
• Authors: Arnav Chavan, Nahush Lele, Udbhav Bamba, Sankalp Dayal, Aditi Raghunathan, Deepak Gupta,
• Abstract summary: Activation outliers in large-scale transformer models pose a fundamental challenge to model quantization.<n>We propose Selective Spectral Decay ($S2D$), a geometrically-principled conditioning method that surgically regularizes only the weight components corresponding to the largest singular values.<n>We show that $S2D$ significantly reduces activation outliers and produces well-conditioned representations that are inherently quantization-friendly.
• Score: 30.145244497712792
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Activation outliers in large-scale transformer models pose a fundamental challenge to model quantization, creating excessively large ranges that cause severe accuracy drops during quantization. We empirically observe that outlier severity intensifies with pre-training scale (e.g., progressing from CLIP to the more extensively trained SigLIP and SigLIP2). Through theoretical analysis as well as empirical correlation studies, we establish the direct link between these activation outliers and dominant singular values of the weights. Building on this insight, we propose Selective Spectral Decay ($S^2D$), a geometrically-principled conditioning method that surgically regularizes only the weight components corresponding to the largest singular values during fine-tuning. Through extensive experiments, we demonstrate that $S^2D$ significantly reduces activation outliers and produces well-conditioned representations that are inherently quantization-friendly. Models trained with $S^2D$ achieve up to 7% improved PTQ accuracy on ImageNet under W4A4 quantization and 4% gains when combined with QAT. These improvements also generalize across downstream tasks and vision-language models, enabling the scaling of increasingly large and rigorously trained models without sacrificing deployment efficiency.

Related papers

• Astro: Activation-guided Structured Regularization for Outlier-Robust LLM Post-Training Quantization [56.5199302532159]
  We propose an Activation-guided Structured Regularization framework to suppress the negative effects of outliers.<n>Astro actively reconstructs intrinsically robust weights, aggressively suppressing weight outliers corresponding to high-magnitude activations.<n>Astro achieves highly competitive performance; notably, on LLaMA-2-7B, it achieves better performance than complex learning-based rotation methods with almost 1/3 of the quantization time.
  arXiv  Detail & Related papers  (2026-02-07T15:50:18Z)
• D$^2$Quant: Accurate Low-bit Post-Training Weight Quantization for LLMs [33.883527341335856]
  Weight-only post-training quantization (PTQ) is appealing as it reduces memory usage and enables practical speedup without low-bit operators or specialized hardware.<n> accuracy often degrades significantly in weight-only PTQ at sub-4-bit precision.<n>We propose D$2$Quant, a novel weight-only PTQ framework that improves quantization from both the weight and activation perspectives.
  arXiv  Detail & Related papers  (2026-01-30T05:49:48Z)
• Extreme Model Compression with Structured Sparsity at Low Precision [10.976782748075067]
  Deep neural networks (DNNs) are used in many applications, but their large size and high computational cost make them hard to run on devices with limited resources.<n>Two widely used techniques to address this challenge are weight quantization, which lowers the precision of all weights, and structured sparsity, which removes unimportant weights while retaining the important ones at full precision.<n>We introduce SLOPE Structured Sparsity at Low Precision, a unified framework, to effectively combine structured sparsity and low-bit quantization in a principled way.
  arXiv  Detail & Related papers  (2025-11-11T15:37:55Z)
• A Convergence Analysis of Adaptive Optimizers under Floating-point Quantization [32.97211471008323]
  We introduce the first theoretical framework of adaptive convergences, including Adam and Muon, under floating-point quantization of gradients, weights, and states.<n>We show that both algorithms retain convergence rates close to their full-precision counterparts provided mantissa length scales only logarithmically with the number of iterations.<n>Our analysis further reveals that Adam is highly sensitive to and second-moment quantization weights due to its reliance on $beta to 1$, while Muon requires weaker error control and is thus potentially more robust.
  arXiv  Detail & Related papers  (2025-10-24T10:16:23Z)
• Beyond Outliers: A Study of Optimizers Under Quantization [82.75879062804955]
  We study impact of choice on model robustness under quantization.<n>We evaluate how model performance degrades when trained with different baselines.<n>We derive scaling laws for quantization-aware training under different parameters.
  arXiv  Detail & Related papers  (2025-09-27T21:15:22Z)
• Improving Quantization with Post-Training Model Expansion [1.6118059133556246]
  We show it is possible to selectively expand the size of a pre-trained large language model to improve model quality without end-to-end retraining.<n>In particular, when quantizing the weights and activations to 4 bits for Llama3 1B, we reduce the gap to full-precision perplexity by an average of 9%.
  arXiv  Detail & Related papers  (2025-03-21T19:56:59Z)
• RoSTE: An Efficient Quantization-Aware Supervised Fine-Tuning Approach for Large Language Models [53.571195477043496]
  We propose an algorithm named Rotated Straight-Through-Estimator (RoSTE)<n>RoSTE combines quantization-aware supervised fine-tuning (QA-SFT) with an adaptive rotation strategy to reduce activation outliers.<n>Our findings reveal that the prediction error is directly proportional to the quantization error of the converged weights, which can be effectively managed through an optimized rotation configuration.
  arXiv  Detail & Related papers  (2025-02-13T06:44:33Z)
• GWQ: Gradient-Aware Weight Quantization for Large Language Models [56.22507677736051]
  Large language models (LLMs) show impressive performance in solving complex language tasks.<n> compressing LLMs to low bits can enable to deploy on resource-constrained devices.<n>We propose gradient-aware weight quantization (GWQ), the first quantization approach for low-bit weight quantization.
  arXiv  Detail & Related papers  (2024-10-30T11:16:04Z)
• FlatQuant: Flatness Matters for LLM Quantization [58.28221892035609]
  We propose FlatQuant, a new post-training quantization approach that enhances the flatness of weights and activations.<n>Our approach identifies optimal affine transformations for each linear layer, calibrated in hours via a lightweight objective.<n>It achieves less than 1% accuracy drop for W4A4 quantization on the LLaMA-3-70B model, surpassing SpinQuant by 7.5%.
  arXiv  Detail & Related papers  (2024-10-12T08:10:28Z)
• Compensate Quantization Errors+: Quantized Models Are Inquisitive Learners [51.32182730502002]
  We introduce Singular-value Diagonal Expansion to refine weight distributions to achieve better quantization alignment.<n>Our plug-and-play weight-quantization methods demonstrate substantial performance improvements over state-of-the-art approaches.
  arXiv  Detail & Related papers  (2024-07-22T09:45:16Z)
• Norm Tweaking: High-performance Low-bit Quantization of Large Language Models [21.855106896725598]
  We introduce a technique called norm tweaking, which can be used as a plugin in current PTQ methods to achieve high precision. Our method demonstrates significant improvements in both weight-only quantization and joint quantization of weights and activations. Our simple and effective approach makes it more practical for real-world applications.
  arXiv  Detail & Related papers  (2023-09-06T06:51:15Z)
• DQ-BART: Efficient Sequence-to-Sequence Model via Joint Distillation and Quantization [75.72231742114951]
  Large-scale pre-trained sequence-to-sequence models like BART and T5 achieve state-of-the-art performance on many generative NLP tasks. These models pose a great challenge in resource-constrained scenarios owing to their large memory requirements and high latency. We propose to jointly distill and quantize the model, where knowledge is transferred from the full-precision teacher model to the quantized and distilled low-precision student model.
  arXiv  Detail & Related papers  (2022-03-21T18:04:25Z)
• Sparse MoEs meet Efficient Ensembles [49.313497379189315]
  We study the interplay of two popular classes of such models: ensembles of neural networks and sparse mixture of experts (sparse MoEs) We present Efficient Ensemble of Experts (E$3$), a scalable and simple ensemble of sparse MoEs that takes the best of both classes of models, while using up to 45% fewer FLOPs than a deep ensemble.
  arXiv  Detail & Related papers  (2021-10-07T11:58:35Z)

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.