@inproceedings{song-etal-2026-intrinsic,
title = "Intrinsic Attention as Navigator: A Non-Generative Retrieval Method for Multi-Hop Queries",
author = "Song, Linwei and
Zhang, Weijie and
Chen, WeiDi and
Lai, Huixia and
Zhang, Shi and
Xiao, Ruliang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1449/",
pages = "30876--30892",
ISBN = "979-8-89176-462-0",
abstract = "For multi-hop queries in retrieval-augmented generation (RAG), it is crucial to accurately obtain evidence pieces. In this study, we propose a training-free method for multi-hop query retrieval, named Intrinsic Navigating-based Non-Generative Retrieval (IN-Retriever). Instead of generating intermediate reasoning text, the method extracts the bridging clue from the internal attention of the embedding model based on the current context and retrieved candidate documents. During iterative clue extraction and document retrieval, IN-Retriever strips away revealed clues from the embedded query vector and then reranks candidate documents in a targeted manner, thereby steering retrieval toward hidden next-hop evidence. In the experiments, we confirm the feasibility and effectiveness of using intrinsic attention signals to guide external multi-hop retrieval. Experimental results show that our method precisely collects sufficient documents to support multi-hop queries and significantly outperforms sparse and dense retrieval methods."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="song-etal-2026-intrinsic">
<titleInfo>
<title>Intrinsic Attention as Navigator: A Non-Generative Retrieval Method for Multi-Hop Queries</title>
</titleInfo>
<name type="personal">
<namePart type="given">Linwei</namePart>
<namePart type="family">Song</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weijie</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">WeiDi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huixia</namePart>
<namePart type="family">Lai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruliang</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>For multi-hop queries in retrieval-augmented generation (RAG), it is crucial to accurately obtain evidence pieces. In this study, we propose a training-free method for multi-hop query retrieval, named Intrinsic Navigating-based Non-Generative Retrieval (IN-Retriever). Instead of generating intermediate reasoning text, the method extracts the bridging clue from the internal attention of the embedding model based on the current context and retrieved candidate documents. During iterative clue extraction and document retrieval, IN-Retriever strips away revealed clues from the embedded query vector and then reranks candidate documents in a targeted manner, thereby steering retrieval toward hidden next-hop evidence. In the experiments, we confirm the feasibility and effectiveness of using intrinsic attention signals to guide external multi-hop retrieval. Experimental results show that our method precisely collects sufficient documents to support multi-hop queries and significantly outperforms sparse and dense retrieval methods.</abstract>
<identifier type="citekey">song-etal-2026-intrinsic</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1449/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>30876</start>
<end>30892</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Intrinsic Attention as Navigator: A Non-Generative Retrieval Method for Multi-Hop Queries
%A Song, Linwei
%A Zhang, Weijie
%A Chen, WeiDi
%A Lai, Huixia
%A Zhang, Shi
%A Xiao, Ruliang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F song-etal-2026-intrinsic
%X For multi-hop queries in retrieval-augmented generation (RAG), it is crucial to accurately obtain evidence pieces. In this study, we propose a training-free method for multi-hop query retrieval, named Intrinsic Navigating-based Non-Generative Retrieval (IN-Retriever). Instead of generating intermediate reasoning text, the method extracts the bridging clue from the internal attention of the embedding model based on the current context and retrieved candidate documents. During iterative clue extraction and document retrieval, IN-Retriever strips away revealed clues from the embedded query vector and then reranks candidate documents in a targeted manner, thereby steering retrieval toward hidden next-hop evidence. In the experiments, we confirm the feasibility and effectiveness of using intrinsic attention signals to guide external multi-hop retrieval. Experimental results show that our method precisely collects sufficient documents to support multi-hop queries and significantly outperforms sparse and dense retrieval methods.
%U https://aclanthology.org/2026.findings-emnlp.1449/
%P 30876-30892
Markdown (Informal)
[Intrinsic Attention as Navigator: A Non-Generative Retrieval Method for Multi-Hop Queries](https://aclanthology.org/2026.findings-emnlp.1449/) (Song et al., Findings 2026)
ACL