<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lin-etal-2025-semantic">
    <titleInfo>
        <title>Semantic Contribution-Aware Adaptive Retrieval for Black-Box Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Qinhong</namePart>
        <namePart type="family">Lin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhongliang</namePart>
        <namePart type="family">Yang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yuang</namePart>
        <namePart type="family">Cai</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dingfu</namePart>
        <namePart type="family">Yu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Xuan</namePart>
        <namePart type="family">Xu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yu</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Linna</namePart>
        <namePart type="family">Zhou</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-11</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Christos</namePart>
            <namePart type="family">Christodoulopoulos</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Tanmoy</namePart>
            <namePart type="family">Chakraborty</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Carolyn</namePart>
            <namePart type="family">Rose</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Violet</namePart>
            <namePart type="family">Peng</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Suzhou, China</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-335-7</identifier>
    </relatedItem>
    <abstract>Retrieval-Augmented Generation (RAG) plays a critical role in mitigating hallucinations and improving factual accuracy for Large Language Models (LLMs). While dynamic retrieval techniques aim to determine retrieval timing and content based on model intrinsic needs, existing approaches struggle to generalize effectively in black-box model scenarios. To address this limitation, we propose the Semantic Contribution-Aware Adaptive Retrieval (SCAAR) framework. SCAAR iteratively leverages the semantic importance of words in upcoming sentences to dynamically adjust retrieval thresholds and filter information, retaining the top-α% most semantically significant words for constructing retrieval queries. We comprehensively evaluate SCAAR against baseline methods across four long-form, knowledge-intensive generation datasets using four models. Our method achieved the highest score on each dataset with GPT-4o. Extensive experiments also analyze the impact of various hyperparameters within the framework. Our results demonstrate SCAAR’s superior or competitive performance, showcasing its ability to effectively detect model retrieval needs and construct efficient retrieval queries for relevant knowledge about problem-solving in black-box scenarios. Our code is available on https://github.com/linqinhong/SAC.</abstract>
    <identifier type="citekey">lin-etal-2025-semantic</identifier>
    <identifier type="doi">10.18653/v1/2025.findings-emnlp.921</identifier>
    <location>
        <url>https://aclanthology.org/2025.findings-emnlp.921/</url>
    </location>
    <part>
        <date>2025-11</date>
        <extent unit="page">
            <start>16979</start>
            <end>16994</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Semantic Contribution-Aware Adaptive Retrieval for Black-Box Models
%A Lin, Qinhong
%A Yang, Zhongliang
%A Cai, Yuang
%A Yu, Dingfu
%A Xu, Xuan
%A Li, Yu
%A Zhou, Linna
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F lin-etal-2025-semantic
%X Retrieval-Augmented Generation (RAG) plays a critical role in mitigating hallucinations and improving factual accuracy for Large Language Models (LLMs). While dynamic retrieval techniques aim to determine retrieval timing and content based on model intrinsic needs, existing approaches struggle to generalize effectively in black-box model scenarios. To address this limitation, we propose the Semantic Contribution-Aware Adaptive Retrieval (SCAAR) framework. SCAAR iteratively leverages the semantic importance of words in upcoming sentences to dynamically adjust retrieval thresholds and filter information, retaining the top-α% most semantically significant words for constructing retrieval queries. We comprehensively evaluate SCAAR against baseline methods across four long-form, knowledge-intensive generation datasets using four models. Our method achieved the highest score on each dataset with GPT-4o. Extensive experiments also analyze the impact of various hyperparameters within the framework. Our results demonstrate SCAAR’s superior or competitive performance, showcasing its ability to effectively detect model retrieval needs and construct efficient retrieval queries for relevant knowledge about problem-solving in black-box scenarios. Our code is available on https://github.com/linqinhong/SAC.
%R 10.18653/v1/2025.findings-emnlp.921
%U https://aclanthology.org/2025.findings-emnlp.921/
%U https://doi.org/10.18653/v1/2025.findings-emnlp.921
%P 16979-16994

Markdown (Informal)

[Semantic Contribution-Aware Adaptive Retrieval for Black-Box Models](https://aclanthology.org/2025.findings-emnlp.921/) (Lin et al., Findings 2025)

• Semantic Contribution-Aware Adaptive Retrieval for Black-Box Models (Lin et al., Findings 2025)

ACL

• Qinhong Lin, Zhongliang Yang, Yuang Cai, Dingfu Yu, Xuan Xu, Yu Li, and Linna Zhou. 2025. Semantic Contribution-Aware Adaptive Retrieval for Black-Box Models. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 16979–16994, Suzhou, China. Association for Computational Linguistics.