Continual Gradient Low-Rank Projection Fine-Tuning for LLMs
Chenxu Wang, Yilin Lyu, Zicheng Sun, Liping Jing
Abstract
Continual fine-tuning of Large Language Models (LLMs) is hampered by the trade-off between efficiency and expressiveness. Low-Rank Adaptation (LoRA) offers efficiency but constrains the model’s ability to learn new tasks and transfer knowledge due to its low-rank nature and reliance on explicit parameter constraints. We propose GORP ( ̲Gradient L ̲Ow ̲Rank ̲Projection) for Continual Learning, a novel training strategy that overcomes these limitations by synergistically combining full and low-rank parameters and jointly updating within a unified low-rank gradient subspace. GORP expands the optimization space while preserving efficiency and mitigating catastrophic forgetting. Extensive experiments on continual learning benchmarks demonstrate GORP’s superior performance compared to existing state-of-the-art approaches. Code is available at https://github.com/Wcxwcxw/GORP.
BibTeX
@inproceedings{wang-etal-2025-continual,
title = "Continual Gradient Low-Rank Projection Fine-Tuning for {LLM}s",
author = "Wang, Chenxu and
Lyu, Yilin and
Sun, Zicheng and
Jing, Liping",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.721/",
doi = "10.18653/v1/2025.acl-long.721",
pages = "14815--14829",
ISBN = "979-8-89176-251-0"
}