Rethinking Long Context Generation from the Continual Learning Perspective

Zeyuan Yang, Fangzhou Xiong, Peng Li, Yang Liu


Abstract
Due to the limited context window, Large Language Models (LLMs) struggle with processing long contexts. Although fine-tuning can extend the context window, it incurs substantial computation costs. In contrast, recent tuning-free approaches reallocate the attention mechanism or incorporate temporary trainable parameters. In this work, by jointly modeling instance-level generation with a limited context window and learning over sequential data, we rethink the long context generation of LLMs from a continual learning perspective. In practice, we inspect existing representative approaches and analyze their synergy with continual learning strategies. Moreover, we integrate these strategies into current approaches to further boost LLMs’ efficiency in processing long contexts. Comprehensive experiments and analysis confirm the feasibility of continual learning insights for improving long-context processing.
Anthology ID:
2025.coling-main.131
Volume:
Proceedings of the 31st International Conference on Computational Linguistics
Month:
January
Year:
2025
Address:
Abu Dhabi, UAE
Editors:
Owen Rambow, Leo Wanner, Marianna Apidianaki, Hend Al-Khalifa, Barbara Di Eugenio, Steven Schockaert
Venue:
COLING
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1922–1933
Language:
URL:
https://aclanthology.org/2025.coling-main.131/
DOI:
Bibkey:
Cite (ACL):
Zeyuan Yang, Fangzhou Xiong, Peng Li, and Yang Liu. 2025. Rethinking Long Context Generation from the Continual Learning Perspective. In Proceedings of the 31st International Conference on Computational Linguistics, pages 1922–1933, Abu Dhabi, UAE. Association for Computational Linguistics.
Cite (Informal):
Rethinking Long Context Generation from the Continual Learning Perspective (Yang et al., COLING 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.coling-main.131.pdf