mlsanityData to 14 Oct 2026

Universal Properties of Activation Sparsity in Modern Large Language Models

arXiv 2509.00454 ↗First posted 30 Aug 2025cs.LGcited by 7

Not on the map. Why?

Abstract (arXiv)

Activation sparsity is an intriguing property of deep neural networks that has been extensively studied in ReLU-based models, due to its advantages for efficiency, robustness, and interpretability. However, methods relying on exact zero activations do not directly apply to modern Large Language Models (LLMs), leading to fragmented, model-specific strategies for LLM activation sparsity and a gap in its general understanding. In this work, we introduce a general framework for evaluating sparsity robustness in contemporary LLMs and conduct a systematic investigation of this phenomenon in their feedforward~(FFN) layers. Our results uncover universal properties of activation sparsity across diverse model families and scales. Importantly, we observe that the potential for effective activation sparsity grows with model size, highlighting its increasing relevance as models scale. Furthermore, we present the first study of activation sparsity in diffusion-based LLMs. Overall, our work provides a comprehensive perspective and practical guidance for harnessing activation sparsity in LLM design and acceleration.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Builds on 60 of its 73 references that are in the map. First posted on arXiv 30 Aug 2025.

Builds on

60 of 73 references.

References · 60

  1. 2017Attention Is All You Needcited by 60,361
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  4. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  5. 2024Qwen2.5 Technical Reportcited by 4,942
  6. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  7. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,100
  8. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  9. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,975
  10. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
  11. 2019PIQA: Reasoning about Physical Commonsense in Natural Languagecited by 2,883
  12. 2019BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questionscited by 2,183
38 more
  1. 2025Gemma 3 Technical Reportcited by 1,824
  2. 2020GLU Variants Improve Transformercited by 1,804
  3. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,420
  4. 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,336
  5. 2016The LAMBADA dataset: Word prediction requiring a broad discourse contextcited by 992
  6. 2025Large Language Diffusion Modelscited by 912
  7. 2017Crowdsourcing Multiple Choice Science Questionscited by 765
  8. 2024DeepSeek LLM: Scaling Open-Source Language Models with Longtermismcited by 642
  9. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertaintycited by 578
  10. 2024Scaling and evaluating sparse autoencoderscited by 553
  11. 2023DeepCache: Accelerating Diffusion Models for Freecited by 359
  12. 2025EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Testcited by 331
  13. 2024OLMoE: Open Mixture-of-Experts Language Modelscited by 301
  14. 2024Massive Activations in Large Language Modelscited by 294
  15. 2023Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Timecited by 252
  16. 2023Structural Pruning for Diffusion Modelscited by 195
  17. 2024Are We Done with MMLU?cited by 194
  18. 2023LLM in a flash: Efficient Large Language Model Inference with Limited Memorycited by 177
  19. 2018Stochastic Activation Pruning for Robust Adversarial Defensecited by 171
  20. 2021MoEfication: Transformer Feed-forward Layers are Mixtures of Expertscited by 152
  21. 2022The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformerscited by 122
  22. 2024Scaling Laws for Precisioncited by 95
  23. 2023ReLU Strikes Back: Exploiting Activation Sparsity in Large Language Modelscited by 94
  24. 2022SGD with Large Step Sizes Learns Sparse Featurescited by 68
  25. 2024ReLU2 Wins: Discovering Efficient Activation Functions for Sparse LLMscited by 67
  26. 2025Overtrained Language Models Are Harder to Fine-Tunecited by 58
  27. 2024Training-Free Activation Sparsity in Large Language Modelscited by 55
  28. 2019How Can We Be So Dense? The Benefits of Using Highly Sparse Representationscited by 45
  29. 2024ProSparse: Introducing and Enhancing Intrinsic Activation Sparsity within Large Language Modelscited by 45
  30. 2013Low-Rank Approximations for Conditional Feedforward Computation in Deep Neural Networkscited by 44
  31. 2024CATS: Contextually-Aware Thresholding for Sparsity in Large Language Modelscited by 41
  32. 2024Turbo Sparse: Achieving LLM SOTA Performance with Minimal Activated Parameterscited by 38
  33. 2017Compressing DMA Engine: Leveraging Activation Sparsity for Training Deep Neural Networkscited by 37
  34. 2023Emergent Modularity in Pre-trained Transformerscited by 37
  35. 2024Token Pruning for Caching Better: 9 Times Acceleration on Stable Diffusion for Freecited by 30
  36. 2024Sparsing Law: Towards Large Language Models with Greater Activation Sparsitycited by 22
  37. 2024Prompt-prompted Adaptive Structured Pruning for Efficient LLM Generationcited by 21
  38. 2024SparseDM: Toward Sparse Efficient Diffusion Modelscited by 18

50 most cited of 60 All 60 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 6.

7 papers.

  1. 2025Query Circuits: Explaining How Language Models Answer User Promptscited by 0
  2. 2026Uncovering Intra-expert Activation Sparsity for Efficient Mixture-of-Expert Model Executioncited by 0
  3. 2026What does FFN compression change downstream? Same-state causal restoration in diffusion language modelscited by 0
  4. 2026Transformer MLP Gate Thresholds Are Couplings to a Carried Reference Directioncited by 0
  5. 2026Model Casting and Low-Parameter Gating: Towards More Sparsely Activated FFNscited by 0
  6. 2026Activation Sparsity with Weight Approximation for Faster LLM Decoding on Offloaded Weightscited by 0
  7. 2026On the Necessity of Attention-FFN Split in Vision Transformerscited by 0