@inproceedings{hao-etal-2025-glitters,
title = "All That Glitters is Not Gold: Improving Robust Retrieval-Augmented Language Models with Fact-Centric Preference Alignment",
author = "Hao, Jia and
Zhang, Chunhong and
Liu, Jiarun and
Zhao, Haiyu and
Zhan, Zhiqiang and
Hu, Zheng",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-acl.588/",
doi = "10.18653/v1/2025.findings-acl.588",
pages = "11277--11292",
ISBN = "979-8-89176-256-5",
abstract = "Retrieval-augmented language model (RALM) relies on retrieved external knowledge to generate responses, resulting in vulnerability in the face of retrieval results with noisy documents. Previous works integrate additional filters or finetune Large Language Models (LLMs) to learn adaptive retrieval to reduce the performance damage of noisy documents. However, prior noise filtering may lead to the loss of crucial information, and these methods do not focus on distracting documents with high semantic relevance, which is the most challenging problem. In this study, we propose a training method for fact-centric preference alignment (FPA) to improve the ability of LLMs to directly extract useful information from noisy retrieval results without prior filtering. Our method performs positive document mining based on factual consistency and uses LLMs self-generated synthetic data as training data without manual annotation. We evaluate our FPA on four question answering benchmarks, and the experimental results demonstrate that our method achieves significant improvement with a small scale of training data."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hao-etal-2025-glitters">
<titleInfo>
<title>All That Glitters is Not Gold: Improving Robust Retrieval-Augmented Language Models with Fact-Centric Preference Alignment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jia</namePart>
<namePart type="family">Hao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chunhong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiarun</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haiyu</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhiqiang</namePart>
<namePart type="family">Zhan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zheng</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-256-5</identifier>
</relatedItem>
<abstract>Retrieval-augmented language model (RALM) relies on retrieved external knowledge to generate responses, resulting in vulnerability in the face of retrieval results with noisy documents. Previous works integrate additional filters or finetune Large Language Models (LLMs) to learn adaptive retrieval to reduce the performance damage of noisy documents. However, prior noise filtering may lead to the loss of crucial information, and these methods do not focus on distracting documents with high semantic relevance, which is the most challenging problem. In this study, we propose a training method for fact-centric preference alignment (FPA) to improve the ability of LLMs to directly extract useful information from noisy retrieval results without prior filtering. Our method performs positive document mining based on factual consistency and uses LLMs self-generated synthetic data as training data without manual annotation. We evaluate our FPA on four question answering benchmarks, and the experimental results demonstrate that our method achieves significant improvement with a small scale of training data.</abstract>
<identifier type="citekey">hao-etal-2025-glitters</identifier>
<identifier type="doi">10.18653/v1/2025.findings-acl.588</identifier>
<location>
<url>https://aclanthology.org/2025.findings-acl.588/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>11277</start>
<end>11292</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T All That Glitters is Not Gold: Improving Robust Retrieval-Augmented Language Models with Fact-Centric Preference Alignment
%A Hao, Jia
%A Zhang, Chunhong
%A Liu, Jiarun
%A Zhao, Haiyu
%A Zhan, Zhiqiang
%A Hu, Zheng
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Findings of the Association for Computational Linguistics: ACL 2025
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-256-5
%F hao-etal-2025-glitters
%X Retrieval-augmented language model (RALM) relies on retrieved external knowledge to generate responses, resulting in vulnerability in the face of retrieval results with noisy documents. Previous works integrate additional filters or finetune Large Language Models (LLMs) to learn adaptive retrieval to reduce the performance damage of noisy documents. However, prior noise filtering may lead to the loss of crucial information, and these methods do not focus on distracting documents with high semantic relevance, which is the most challenging problem. In this study, we propose a training method for fact-centric preference alignment (FPA) to improve the ability of LLMs to directly extract useful information from noisy retrieval results without prior filtering. Our method performs positive document mining based on factual consistency and uses LLMs self-generated synthetic data as training data without manual annotation. We evaluate our FPA on four question answering benchmarks, and the experimental results demonstrate that our method achieves significant improvement with a small scale of training data.
%R 10.18653/v1/2025.findings-acl.588
%U https://aclanthology.org/2025.findings-acl.588/
%U https://doi.org/10.18653/v1/2025.findings-acl.588
%P 11277-11292
Markdown (Informal)
[All That Glitters is Not Gold: Improving Robust Retrieval-Augmented Language Models with Fact-Centric Preference Alignment](https://aclanthology.org/2025.findings-acl.588/) (Hao et al., Findings 2025)
ACL