@inproceedings{wang-etal-2026-task,
title = "Task-Related In-Context Learning",
author = "Wang, Wenqiang and
Chen, Peng and
Xiao, Yan and
Zhang, Yangshijie and
Lu, Xiaoyue and
Huang, Jianjie and
Cao, Xiaochun",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.2154/",
pages = "43378--43400",
ISBN = "979-8-89176-395-1",
abstract = "Standard in-context learning (ICL) assumes identical output spaces between test and retrieval datasets (fully aligned). However, in practice, these datasets can be fully aligned, partially aligned, or fully disjoint in label space (Output space), forming an information continuum from rich to scarce. Naive ICL often becomes ineffective under such mismatches. In this work, we challenge this assumption by demonstrating that the retrieval dataset need not perfectly align with the test dataset, as long as it remains related to the target task. We propose Task-Related In-Context Learning (TRICL), a unified framework for ICL under output-space mismatch, designed to cover the full continuum of scenarios. TRICL first identifies demonstrations in the mismatched retrieval dataset that are relevant to the test label space via a lightweight Bayesian probabilistic criterion, and uses them to form a related dataset. TRICL then perform ICL on the related dataset to obtain preliminary predictions; finally, TRICL leverage these intermediate predictions to reduce and transform the output space of the original test task, thereby improving the performance of LLMs. Even in the most information-scarce fully disjoint scenario, as long as the retrieval dataset is task-related to the test task, TRICL achieves state-of-the-art (SOTA) results across three LLMs, three task types, and four datasets. Moreover, TRICL remains effective in the fully aligned and partially aligned scenarios, consistently yielding strong gains over competitive baselines. Moreover, TRICL also extends to generative task."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-task">
<titleInfo>
<title>Task-Related In-Context Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wenqiang</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peng</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yan</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yangshijie</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoyue</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianjie</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaochun</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Standard in-context learning (ICL) assumes identical output spaces between test and retrieval datasets (fully aligned). However, in practice, these datasets can be fully aligned, partially aligned, or fully disjoint in label space (Output space), forming an information continuum from rich to scarce. Naive ICL often becomes ineffective under such mismatches. In this work, we challenge this assumption by demonstrating that the retrieval dataset need not perfectly align with the test dataset, as long as it remains related to the target task. We propose Task-Related In-Context Learning (TRICL), a unified framework for ICL under output-space mismatch, designed to cover the full continuum of scenarios. TRICL first identifies demonstrations in the mismatched retrieval dataset that are relevant to the test label space via a lightweight Bayesian probabilistic criterion, and uses them to form a related dataset. TRICL then perform ICL on the related dataset to obtain preliminary predictions; finally, TRICL leverage these intermediate predictions to reduce and transform the output space of the original test task, thereby improving the performance of LLMs. Even in the most information-scarce fully disjoint scenario, as long as the retrieval dataset is task-related to the test task, TRICL achieves state-of-the-art (SOTA) results across three LLMs, three task types, and four datasets. Moreover, TRICL remains effective in the fully aligned and partially aligned scenarios, consistently yielding strong gains over competitive baselines. Moreover, TRICL also extends to generative task.</abstract>
<identifier type="citekey">wang-etal-2026-task</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.2154/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>43378</start>
<end>43400</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Task-Related In-Context Learning
%A Wang, Wenqiang
%A Chen, Peng
%A Xiao, Yan
%A Zhang, Yangshijie
%A Lu, Xiaoyue
%A Huang, Jianjie
%A Cao, Xiaochun
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F wang-etal-2026-task
%X Standard in-context learning (ICL) assumes identical output spaces between test and retrieval datasets (fully aligned). However, in practice, these datasets can be fully aligned, partially aligned, or fully disjoint in label space (Output space), forming an information continuum from rich to scarce. Naive ICL often becomes ineffective under such mismatches. In this work, we challenge this assumption by demonstrating that the retrieval dataset need not perfectly align with the test dataset, as long as it remains related to the target task. We propose Task-Related In-Context Learning (TRICL), a unified framework for ICL under output-space mismatch, designed to cover the full continuum of scenarios. TRICL first identifies demonstrations in the mismatched retrieval dataset that are relevant to the test label space via a lightweight Bayesian probabilistic criterion, and uses them to form a related dataset. TRICL then perform ICL on the related dataset to obtain preliminary predictions; finally, TRICL leverage these intermediate predictions to reduce and transform the output space of the original test task, thereby improving the performance of LLMs. Even in the most information-scarce fully disjoint scenario, as long as the retrieval dataset is task-related to the test task, TRICL achieves state-of-the-art (SOTA) results across three LLMs, three task types, and four datasets. Moreover, TRICL remains effective in the fully aligned and partially aligned scenarios, consistently yielding strong gains over competitive baselines. Moreover, TRICL also extends to generative task.
%U https://aclanthology.org/2026.findings-acl.2154/
%P 43378-43400
Markdown (Informal)
[Task-Related In-Context Learning](https://aclanthology.org/2026.findings-acl.2154/) (Wang et al., Findings 2026)
ACL
- Wenqiang Wang, Peng Chen, Yan Xiao, Yangshijie Zhang, Xiaoyue Lu, Jianjie Huang, and Xiaochun Cao. 2026. Task-Related In-Context Learning. In Findings of the Association for Computational Linguistics: ACL 2026, pages 43378–43400, San Diego, California, United States. Association for Computational Linguistics.