@inproceedings{zhang-etal-2026-training,
title = "Training-Free Generative Retrieval",
author = "Zhang, Yingchen and
Zhang, Ruqing and
Guo, Jiafeng and
Rijke, Maarten de and
Lv, Fuyu and
Cheng, Xueqi",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2202/",
pages = "50054--50069",
ISBN = "979-8-89176-459-0",
abstract = "Large language models (LLMs) are emerging as versatile foundations for generative retrieval (GR), a paradigm that directly generates relevant document identifiers (docids) given a query. Existing ways of using LLMs for GR typically rely on heavy post-hoc training to fit a relevance data prior, often leading to overfitting and poor cross-domain generalization. Can retrieval, like other language tasks, be performed without retriever-specific training using LLMs? We conduct a preliminary study and find that constrained decoding, required to ensure valid docids, often misaligns with the model{'}s natural free-generation patterns. We propose FreeGR, a plug-and-play framework that adapts retrieval data to a LLM{'}s habitual generation behavior. FreeGR consists of three components: an LLM self-annotated docid generator, an LLM-preferred input context designer, and two decoding strategies that trade off retrieval effectiveness and efficiency. It requires no retriever-specific parameter optimization, is model-agnostic, and scales across LLM families and sizes while preserving their generative capabilities. Experiments on the NQ, MS MARCO, and BEIR datasets show that FreeGR achieves competitive retrieval effectiveness compared with several trained generative and dense retrievers, without retriever-specific parameter optimization."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-training">
<titleInfo>
<title>Training-Free Generative Retrieval</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yingchen</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruqing</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiafeng</namePart>
<namePart type="family">Guo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Maarten</namePart>
<namePart type="given">de</namePart>
<namePart type="family">Rijke</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fuyu</namePart>
<namePart type="family">Lv</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xueqi</namePart>
<namePart type="family">Cheng</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Large language models (LLMs) are emerging as versatile foundations for generative retrieval (GR), a paradigm that directly generates relevant document identifiers (docids) given a query. Existing ways of using LLMs for GR typically rely on heavy post-hoc training to fit a relevance data prior, often leading to overfitting and poor cross-domain generalization. Can retrieval, like other language tasks, be performed without retriever-specific training using LLMs? We conduct a preliminary study and find that constrained decoding, required to ensure valid docids, often misaligns with the model’s natural free-generation patterns. We propose FreeGR, a plug-and-play framework that adapts retrieval data to a LLM’s habitual generation behavior. FreeGR consists of three components: an LLM self-annotated docid generator, an LLM-preferred input context designer, and two decoding strategies that trade off retrieval effectiveness and efficiency. It requires no retriever-specific parameter optimization, is model-agnostic, and scales across LLM families and sizes while preserving their generative capabilities. Experiments on the NQ, MS MARCO, and BEIR datasets show that FreeGR achieves competitive retrieval effectiveness compared with several trained generative and dense retrievers, without retriever-specific parameter optimization.</abstract>
<identifier type="citekey">zhang-etal-2026-training</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2202/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>50054</start>
<end>50069</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Training-Free Generative Retrieval
%A Zhang, Yingchen
%A Zhang, Ruqing
%A Guo, Jiafeng
%A Rijke, Maarten de
%A Lv, Fuyu
%A Cheng, Xueqi
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhang-etal-2026-training
%X Large language models (LLMs) are emerging as versatile foundations for generative retrieval (GR), a paradigm that directly generates relevant document identifiers (docids) given a query. Existing ways of using LLMs for GR typically rely on heavy post-hoc training to fit a relevance data prior, often leading to overfitting and poor cross-domain generalization. Can retrieval, like other language tasks, be performed without retriever-specific training using LLMs? We conduct a preliminary study and find that constrained decoding, required to ensure valid docids, often misaligns with the model’s natural free-generation patterns. We propose FreeGR, a plug-and-play framework that adapts retrieval data to a LLM’s habitual generation behavior. FreeGR consists of three components: an LLM self-annotated docid generator, an LLM-preferred input context designer, and two decoding strategies that trade off retrieval effectiveness and efficiency. It requires no retriever-specific parameter optimization, is model-agnostic, and scales across LLM families and sizes while preserving their generative capabilities. Experiments on the NQ, MS MARCO, and BEIR datasets show that FreeGR achieves competitive retrieval effectiveness compared with several trained generative and dense retrievers, without retriever-specific parameter optimization.
%U https://aclanthology.org/2026.emnlp-main.2202/
%P 50054-50069
Markdown (Informal)
[Training-Free Generative Retrieval](https://aclanthology.org/2026.emnlp-main.2202/) (Zhang et al., EMNLP 2026)
ACL
- Yingchen Zhang, Ruqing Zhang, Jiafeng Guo, Maarten de Rijke, Fuyu Lv, and Xueqi Cheng. 2026. Training-Free Generative Retrieval. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 50054–50069, Budapest, Hungary. Association for Computational Linguistics.