@article{li-etal-2025-keft,
title = "{KEFT}: Knowledge-Enhanced Fine-Tuning for Large Language Models in Domain-Specific Question Answering",
author = "Li, Haiyun and
Zhang, Jixin and
Shen, Hua and
Cheng, Ke and
Huang, Xiaofeng",
journal = "Transactions of the Association for Computational Linguistics",
volume = "13",
year = "2025",
address = "Cambridge, MA",
publisher = "MIT Press",
url = "https://aclanthology.org/2025.tacl-1.49/",
doi = "10.1162/tacl.a.31",
pages = "1056--1067",
abstract = "The rapid advancement of large language models (LLMs) has opened up promising opportunities for their downstream applications in question-answering (QA), such as ChatGPT, ChatGLM, etc. However, such LLMs do not perform very well in domain-specific QA tasks without fine-tuning. But directly fine-tuning LLMs on domain-specific corpus data may lead to catastrophic forgetting, causing the LLMs to lose their general language capability. To address this problem, we propose the Knowledge-Enhanced Fine-Tuning (KEFT) method, an unsupervised fine-tuning approach to enhance the knowledge capability of LLMs in domain-specific QA tasks while preserving their general language capability. KEFT leverages the inherent language comprehension of pre-trained LLMs to generate synthetic-QA datasets from domain-specific corpus data autonomously for fine-tuning, and adopts a Low-Rank Adaptation (LoRA) method to further alleviate over-fitting. Furthermore, to enhance the representation of domain-specific knowledge, we introduce a knowledge-enhanced fine-tuning loss function, which encourages the model to learn the knowledge-question connection, thereby generating natural and knowledgeable answers. Our evaluations across multiple domain-specific datasets demonstrate that KEFT surpasses state-of-the-art fine-tuning approaches, enhancing the performance of various LLMs in QA tasks in both English and Chinese languages."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2025-keft">
<titleInfo>
<title>KEFT: Knowledge-Enhanced Fine-Tuning for Large Language Models in Domain-Specific Question Answering</title>
</titleInfo>
<name type="personal">
<namePart type="given">Haiyun</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jixin</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hua</namePart>
<namePart type="family">Shen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ke</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaofeng</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<genre authority="bibutilsgt">journal article</genre>
<relatedItem type="host">
<titleInfo>
<title>Transactions of the Association for Computational Linguistics</title>
</titleInfo>
<originInfo>
<issuance>continuing</issuance>
<publisher>MIT Press</publisher>
<place>
<placeTerm type="text">Cambridge, MA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">periodical</genre>
<genre authority="bibutilsgt">academic journal</genre>
</relatedItem>
<abstract>The rapid advancement of large language models (LLMs) has opened up promising opportunities for their downstream applications in question-answering (QA), such as ChatGPT, ChatGLM, etc. However, such LLMs do not perform very well in domain-specific QA tasks without fine-tuning. But directly fine-tuning LLMs on domain-specific corpus data may lead to catastrophic forgetting, causing the LLMs to lose their general language capability. To address this problem, we propose the Knowledge-Enhanced Fine-Tuning (KEFT) method, an unsupervised fine-tuning approach to enhance the knowledge capability of LLMs in domain-specific QA tasks while preserving their general language capability. KEFT leverages the inherent language comprehension of pre-trained LLMs to generate synthetic-QA datasets from domain-specific corpus data autonomously for fine-tuning, and adopts a Low-Rank Adaptation (LoRA) method to further alleviate over-fitting. Furthermore, to enhance the representation of domain-specific knowledge, we introduce a knowledge-enhanced fine-tuning loss function, which encourages the model to learn the knowledge-question connection, thereby generating natural and knowledgeable answers. Our evaluations across multiple domain-specific datasets demonstrate that KEFT surpasses state-of-the-art fine-tuning approaches, enhancing the performance of various LLMs in QA tasks in both English and Chinese languages.</abstract>
<identifier type="citekey">li-etal-2025-keft</identifier>
<identifier type="doi">10.1162/tacl.a.31</identifier>
<location>
<url>https://aclanthology.org/2025.tacl-1.49/</url>
</location>
<part>
<date>2025</date>
<detail type="volume"><number>13</number></detail>
<extent unit="page">
<start>1056</start>
<end>1067</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Journal Article
%T KEFT: Knowledge-Enhanced Fine-Tuning for Large Language Models in Domain-Specific Question Answering
%A Li, Haiyun
%A Zhang, Jixin
%A Shen, Hua
%A Cheng, Ke
%A Huang, Xiaofeng
%J Transactions of the Association for Computational Linguistics
%D 2025
%V 13
%I MIT Press
%C Cambridge, MA
%F li-etal-2025-keft
%X The rapid advancement of large language models (LLMs) has opened up promising opportunities for their downstream applications in question-answering (QA), such as ChatGPT, ChatGLM, etc. However, such LLMs do not perform very well in domain-specific QA tasks without fine-tuning. But directly fine-tuning LLMs on domain-specific corpus data may lead to catastrophic forgetting, causing the LLMs to lose their general language capability. To address this problem, we propose the Knowledge-Enhanced Fine-Tuning (KEFT) method, an unsupervised fine-tuning approach to enhance the knowledge capability of LLMs in domain-specific QA tasks while preserving their general language capability. KEFT leverages the inherent language comprehension of pre-trained LLMs to generate synthetic-QA datasets from domain-specific corpus data autonomously for fine-tuning, and adopts a Low-Rank Adaptation (LoRA) method to further alleviate over-fitting. Furthermore, to enhance the representation of domain-specific knowledge, we introduce a knowledge-enhanced fine-tuning loss function, which encourages the model to learn the knowledge-question connection, thereby generating natural and knowledgeable answers. Our evaluations across multiple domain-specific datasets demonstrate that KEFT surpasses state-of-the-art fine-tuning approaches, enhancing the performance of various LLMs in QA tasks in both English and Chinese languages.
%R 10.1162/tacl.a.31
%U https://aclanthology.org/2025.tacl-1.49/
%U https://doi.org/10.1162/tacl.a.31
%P 1056-1067
Markdown (Informal)
[KEFT: Knowledge-Enhanced Fine-Tuning for Large Language Models in Domain-Specific Question Answering](https://aclanthology.org/2025.tacl-1.49/) (Li et al., TACL 2025)
ACL