# Activation Sparsity Opportunities for Compressing General Large Language Models

arXiv: https://arxiv.org/abs/2412.12178 · HTML: https://mlsanity.com/paper/2412.12178/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 35 references that are in the map.
- First posted on arXiv 13 Dec 2024.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [Knowledge Distillation: A Survey](https://mlsanity.com/paper/2006.05525/) (2020)
- [Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference](https://mlsanity.com/paper/1712.05877/) (2017)
- [Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning](https://mlsanity.com/paper/1702.03118/) (2017)
- [The Falcon Series of Open Language Models](https://mlsanity.com/paper/2311.16867/) (2023)
- [Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Time](https://mlsanity.com/paper/2310.17157/) (2023)
- [LLM in a flash: Efficient Large Language Model Inference with Limited Memory](https://mlsanity.com/paper/2312.11514/) (2023)
- [Distilling Task-Specific Knowledge from BERT into Simple Neural Networks](https://mlsanity.com/paper/1903.12136/) (2019)
- [Quantization Networks](https://mlsanity.com/paper/1911.09464/) (2019)
- [MoEfication: Transformer Feed-forward Layers are Mixtures of Experts](https://mlsanity.com/paper/2110.01786/) (2021)
- [The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformers](https://mlsanity.com/paper/2210.06313/) (2022)
- [ReLU2 Wins: Discovering Efficient Activation Functions for Sparse LLMs](https://mlsanity.com/paper/2402.03804/) (2024)
- [One Wide Feedforward is All You Need](https://mlsanity.com/paper/2309.01826/) (2023)
- [ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models](https://mlsanity.com/paper/2406.16635/) (2024)
- [The Role of LLMs in Sustainable Smart Cities: Applications, Challenges, and Future Directions](https://mlsanity.com/paper/2402.14596/) (2024)
- [GraphChallenge.org Sparse Deep Neural Network Performance](https://mlsanity.com/paper/2004.01181/) (2020)
- [FL-Tuning: Layer Tuning for Feed-Forward Network in Transformer](https://mlsanity.com/paper/2206.15312/) (2022)

## Built on by (most cited first, 1 of 1)

- [Unifying Depth and Width Pruning for LLMs via Binary Knapsack Optimization](https://mlsanity.com/paper/2608.12953/) (2026)

## Abstract (arXiv)

Deploying local AI models, such as Large Language Models (LLMs), to edge devices can substantially enhance devices' independent capabilities, alleviate the server's burden, and lower the response time. Owing to these tremendous potentials, many big tech companies have released several lightweight Small Language Models (SLMs) to bridge this gap. However, we still have huge motivations to deploy more powerful (LLMs) AI models on edge devices and enhance their smartness level. Unlike the conventional approaches for AI model compression, we investigate activation sparsity. The activation sparsity method is orthogonal and combinable with existing techniques to maximize the compression rate while maintaining great accuracy. LLMs' Feed-Forward Network (FFN) components, which typically comprise a large proportion of parameters (around 2/3), ensure that our FFN optimizations would have a better chance of achieving effective compression. Moreover, our findings are beneficial to general LLMs and are not restricted to ReLU-based models. This work systematically investigates the tradeoff between enforcing activation sparsity and perplexity (accuracy) on state-of-the-art LLMs. Our empirical analysis demonstrates that we can obtain around 50% of main memory and computing reductions for critical FFN components with negligible accuracy degradation. This extra 50% sparsity does not naturally exist in the current LLMs, which require tuning LLMs' activation outputs by injecting zero-enforcing thresholds. To obtain the benefits of activation sparsity, we provide a guideline for the system architect for LLM prediction and prefetching. The success prediction allows the system to prefetch the necessary weights while omitting the inactive ones and their successors, therefore lowering cache and memory pollution and reducing LLM execution time on resource-constrained edge devices.