paper-with-me

홈 › Papers

MOYU: A Theoretical Study on Massive Over-activation Yielded Uplifts in LLMs

2024-06-18 · Chi Ma, Mincong Huang, Chao Wang, Yujie Wang, Lei Yu

Massive Over-activation Yielded Uplifts(MOYU) is an inherent property of large language models, and dynamic activation(DA) based on the MOYU property is a clever yet under-explored strategy designed to accelerate inference in these models. Existing methods that utilize MOYU often face a significant 'Impossible Trinity': struggling to simultaneously maintain model performance, enhance inference speed, and extend applicability across various architectures. Due to the theoretical ambiguities surrounding MOYU, this paper elucidates the root cause of the MOYU property and outlines the mechanisms behind two primary limitations encountered by current DA methods: 1) history-related activation uncertainty, and 2) semantic-irrelevant activation inertia. Our analysis not only underscores the limitations of current dynamic activation strategies within large-scale LLaMA models but also proposes opportunities for refining the design of future sparsity schemes.

📄 PDF Abstract BibTeX arXiv:2406.12569

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

LLaMA LLaMA is a collection of foundation language models ranging from 7B to 65B parameters. It is based on the transformer architecture with various improvements that were…

Similar Papers 제목 키워드 기반

Moyun: A Diffusion-Based Model for Style-Specific Chinese Calligraphy Generation

2024-10-10 · Kaiyuan Liu, Jiahao Mei, Hengyu Zhang, Yihuai Zhang 외

Although Chinese calligraphy generation has achieved style transfer, generating calligraphy by specifying the calligrapher, font, and character style remains challenging. To address this, we propose a new Chinese calligr…

MambaStyle Transfer

Attention Sinks Induce Gradient Sinks: Massive Activations as Gradient Regulators in Transformers

2026-03-18 · Yihong Chen, Zhouchen Lin, Quanming Yao arxiv

Attention sinks and massive activations are recurring and closely related phenomena in Transformer models. Existing explanations have largely focused on the forward pass, yet in pre-norm Transformers, large residual-stre…

Massive Activations in Large Language Models

2024-02-27 · MingJie Sun, Xinlei Chen, J. Zico Kolter, Zhuang Liu

We observe an empirical phenomenon in Large Language Models (LLMs) -- very few activations exhibit significantly larger values than others (e.g., 100,000 times larger). We call them massive activations. First, we demonst…

Attention Sinks and Compression Valleys in LLMs are Two Sides of the Same Coin

2025-10-07 · Enrique Queipo-de-Llano, Álvaro Arroyo, Federico Barbero, Xiaowen Dong 외 arxiv

Attention sinks and compression valleys have attracted significant attention as two puzzling phenomena in large language models, but have been studied in isolation. In this work, we present a surprising connection betwee…

Mitigating Attention Sinks and Massive Activations in Audio-Visual Speech Recognition with LLMs

2025-10-26 · Anand, Umberto Cappellazzo, Stavros Petridis, Maja Pantic arxiv

Large language models (LLMs) have recently advanced auditory speech recognition (ASR), visual speech recognition (VSR), and audio-visual speech recognition (AVSR). However, understanding of their internal dynamics under …

Audio-Visual Speech Recognition