@inproceedings{guan-etal-2025-multi,
title = "Multi-Stage {LLM} Fine-Tuning with a Continual Learning Setting",
author = "Guan, Changhao and
Huang, Chao and
Li, Hongliang and
Li, You and
Cheng, Ning and
Liu, Zihe and
Chen, Yufeng and
Xu, Jinan and
Liu, Jian",
editor = "Chiruzzo, Luis and
Ritter, Alan and
Wang, Lu",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2025",
month = apr,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-naacl.303/",
doi = "10.18653/v1/2025.findings-naacl.303",
pages = "5484--5498",
ISBN = "979-8-89176-195-7",
abstract = "In recent years, large language models (LLMs) have made significant progress in knowledge-intensive applications. However, when adapting them to specific domains, we may encounter a multi-stage continuous learning scenario, especially in cases where domain knowledge evolves rapidly.This issue severely limits traditional fine-tuning approaches for LLMs.To overcome this limitation, we propose a new learning paradigm designed specifically for multi-stage continuous learning. This paradigm includes a preference-based learning bias to identify potential knowledge conflicts, as well as a self-distillation-based data augmentation strategy to expand and enrich the training corpus, thereby improving the integration of knowledge-compatible information.In the experiments, we show that our proposed method achieves a significant improvement in accuracy after 7 stages of fine-tuning compared to previous methods, while also demonstrating excellent performance in preserving general knowledge.We have released our code and dataset at Multi-Stage-Learning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="guan-etal-2025-multi">
<titleInfo>
<title>Multi-Stage LLM Fine-Tuning with a Continual Learning Setting</title>
</titleInfo>
<name type="personal">
<namePart type="given">Changhao</namePart>
<namePart type="family">Guan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chao</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hongliang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">You</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ning</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zihe</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yufeng</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinan</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jian</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: NAACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Luis</namePart>
<namePart type="family">Chiruzzo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="family">Ritter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-195-7</identifier>
</relatedItem>
<abstract>In recent years, large language models (LLMs) have made significant progress in knowledge-intensive applications. However, when adapting them to specific domains, we may encounter a multi-stage continuous learning scenario, especially in cases where domain knowledge evolves rapidly.This issue severely limits traditional fine-tuning approaches for LLMs.To overcome this limitation, we propose a new learning paradigm designed specifically for multi-stage continuous learning. This paradigm includes a preference-based learning bias to identify potential knowledge conflicts, as well as a self-distillation-based data augmentation strategy to expand and enrich the training corpus, thereby improving the integration of knowledge-compatible information.In the experiments, we show that our proposed method achieves a significant improvement in accuracy after 7 stages of fine-tuning compared to previous methods, while also demonstrating excellent performance in preserving general knowledge.We have released our code and dataset at Multi-Stage-Learning.</abstract>
<identifier type="citekey">guan-etal-2025-multi</identifier>
<identifier type="doi">10.18653/v1/2025.findings-naacl.303</identifier>
<location>
<url>https://aclanthology.org/2025.findings-naacl.303/</url>
</location>
<part>
<date>2025-04</date>
<extent unit="page">
<start>5484</start>
<end>5498</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Multi-Stage LLM Fine-Tuning with a Continual Learning Setting
%A Guan, Changhao
%A Huang, Chao
%A Li, Hongliang
%A Li, You
%A Cheng, Ning
%A Liu, Zihe
%A Chen, Yufeng
%A Xu, Jinan
%A Liu, Jian
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Findings of the Association for Computational Linguistics: NAACL 2025
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-195-7
%F guan-etal-2025-multi
%X In recent years, large language models (LLMs) have made significant progress in knowledge-intensive applications. However, when adapting them to specific domains, we may encounter a multi-stage continuous learning scenario, especially in cases where domain knowledge evolves rapidly.This issue severely limits traditional fine-tuning approaches for LLMs.To overcome this limitation, we propose a new learning paradigm designed specifically for multi-stage continuous learning. This paradigm includes a preference-based learning bias to identify potential knowledge conflicts, as well as a self-distillation-based data augmentation strategy to expand and enrich the training corpus, thereby improving the integration of knowledge-compatible information.In the experiments, we show that our proposed method achieves a significant improvement in accuracy after 7 stages of fine-tuning compared to previous methods, while also demonstrating excellent performance in preserving general knowledge.We have released our code and dataset at Multi-Stage-Learning.
%R 10.18653/v1/2025.findings-naacl.303
%U https://aclanthology.org/2025.findings-naacl.303/
%U https://doi.org/10.18653/v1/2025.findings-naacl.303
%P 5484-5498
Markdown (Informal)
[Multi-Stage LLM Fine-Tuning with a Continual Learning Setting](https://aclanthology.org/2025.findings-naacl.303/) (Guan et al., Findings 2025)
ACL
- Changhao Guan, Chao Huang, Hongliang Li, You Li, Ning Cheng, Zihe Liu, Yufeng Chen, Jinan Xu, and Jian Liu. 2025. Multi-Stage LLM Fine-Tuning with a Continual Learning Setting. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 5484–5498, Albuquerque, New Mexico. Association for Computational Linguistics.