CoLA: Compute-Efficient Pre-Training of LLMs via Low-Rank Activation

Ziyue Liu, Ruijie Zhang, Zhengyang Wang, Mingsong Yan, Zi Yang, Paul D. Hovland, Bogdan Nicolae, Franck Cappello, Sui Tang, Zheng Zhang


Abstract
The full-size MLPs and the projection layers in attention introduce tremendous model sizes of large language models (LLMs), consuming extensive computational resources in pre-training. We empirically observe that the activations of pre-trained LLMs exhibit low-rank property. Motivated by such observations, we propose **CoLA** and its memory-efficient implementation, **CoLA-M**, to replace these full-size layers with compute-efficient **auto-encoders** that naturally enforce low-rank activations throughout training. This fundamental architectural change eliminates the activation redundancy and significantly boosts model capacity and training efficiency. Experiments on LLaMA models with 60 million to 7 billion parameters show that CoLA reduces the computing cost by 2\pmb{\times} and improves training throughput by 1.86\pmb{\times} while maintaining full-rank level performance. CoLA-M further squeezes memory cost without sacrificing throughput, offering a pre-training approach with collectively superior parameter, computing, and memory efficiency. The LLMs produced are also 2\pmb{\times} smaller, enabling faster inference with lower memory cost on resource-constrained platforms.
Anthology ID:
2025.emnlp-main.230
Volume:
Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2025
Address:
Suzhou, China
Editors:
Christos Christodoulopoulos, Tanmoy Chakraborty, Carolyn Rose, Violet Peng
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4627–4645
Language:
URL:
https://aclanthology.org/2025.emnlp-main.230/
DOI:
Bibkey:
Cite (ACL):
Ziyue Liu, Ruijie Zhang, Zhengyang Wang, Mingsong Yan, Zi Yang, Paul D. Hovland, Bogdan Nicolae, Franck Cappello, Sui Tang, and Zheng Zhang. 2025. CoLA: Compute-Efficient Pre-Training of LLMs via Low-Rank Activation. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 4627–4645, Suzhou, China. Association for Computational Linguistics.
Cite (Informal):
CoLA: Compute-Efficient Pre-Training of LLMs via Low-Rank Activation (Liu et al., EMNLP 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.emnlp-main.230.pdf
Checklist:
 2025.emnlp-main.230.checklist.pdf