<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="milios-etal-2023-context">
    <titleInfo>
        <title>In-Context Learning for Text Classification with Many Labels</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Aristides</namePart>
        <namePart type="family">Milios</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Siva</namePart>
        <namePart type="family">Reddy</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dzmitry</namePart>
        <namePart type="family">Bahdanau</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2023-12</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 1st GenBench Workshop on (Benchmarking) Generalisation in NLP</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Dieuwke</namePart>
            <namePart type="family">Hupkes</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Verna</namePart>
            <namePart type="family">Dankers</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Khuyagbaatar</namePart>
            <namePart type="family">Batsuren</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Koustuv</namePart>
            <namePart type="family">Sinha</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Amirhossein</namePart>
            <namePart type="family">Kazemnejad</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <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">Ryan</namePart>
            <namePart type="family">Cotterell</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Elia</namePart>
            <namePart type="family">Bruni</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Singapore</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>In-context learning (ICL) using large language models for tasks with many labels is challenging due to the limited context window, which makes it difficult to fit a sufficient number of examples in the prompt. In this paper, we use a pre-trained dense retrieval model to bypass this limitation, giving the model only a partial view of the full label space for each inference call. Testing with recent open-source LLMs (OPT, LLaMA), we set new state of the art performance in few-shot settings for three common intent classification datasets, with no fine-tuning. We also surpass fine-tuned performance on fine-grained sentiment classification in certain cases. We analyze the performance across number of in-context examples and different model scales, showing that larger models are necessary to effectively make use of larger context lengths for ICL. By running several ablations, we analyze the model‘s use of: a) the similarity of the in-context examples to the current input, b) the semantic content of the class names, and c) the correct correspondence between examples and labels. We demonstrate that all three are needed to varying degrees depending on the domain, contrary to certain recent works.</abstract>
    <identifier type="citekey">milios-etal-2023-context</identifier>
    <identifier type="doi">10.18653/v1/2023.genbench-1.14</identifier>
    <location>
        <url>https://aclanthology.org/2023.genbench-1.14/</url>
    </location>
    <part>
        <date>2023-12</date>
        <extent unit="page">
            <start>173</start>
            <end>184</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T In-Context Learning for Text Classification with Many Labels
%A Milios, Aristides
%A Reddy, Siva
%A Bahdanau, Dzmitry
%Y Hupkes, Dieuwke
%Y Dankers, Verna
%Y Batsuren, Khuyagbaatar
%Y Sinha, Koustuv
%Y Kazemnejad, Amirhossein
%Y Christodoulopoulos, Christos
%Y Cotterell, Ryan
%Y Bruni, Elia
%S Proceedings of the 1st GenBench Workshop on (Benchmarking) Generalisation in NLP
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F milios-etal-2023-context
%X In-context learning (ICL) using large language models for tasks with many labels is challenging due to the limited context window, which makes it difficult to fit a sufficient number of examples in the prompt. In this paper, we use a pre-trained dense retrieval model to bypass this limitation, giving the model only a partial view of the full label space for each inference call. Testing with recent open-source LLMs (OPT, LLaMA), we set new state of the art performance in few-shot settings for three common intent classification datasets, with no fine-tuning. We also surpass fine-tuned performance on fine-grained sentiment classification in certain cases. We analyze the performance across number of in-context examples and different model scales, showing that larger models are necessary to effectively make use of larger context lengths for ICL. By running several ablations, we analyze the model‘s use of: a) the similarity of the in-context examples to the current input, b) the semantic content of the class names, and c) the correct correspondence between examples and labels. We demonstrate that all three are needed to varying degrees depending on the domain, contrary to certain recent works.
%R 10.18653/v1/2023.genbench-1.14
%U https://aclanthology.org/2023.genbench-1.14/
%U https://doi.org/10.18653/v1/2023.genbench-1.14
%P 173-184

Markdown (Informal)

[In-Context Learning for Text Classification with Many Labels](https://aclanthology.org/2023.genbench-1.14/) (Milios et al., GenBench 2023)

• In-Context Learning for Text Classification with Many Labels (Milios et al., GenBench 2023)

ACL

• Aristides Milios, Siva Reddy, and Dzmitry Bahdanau. 2023. In-Context Learning for Text Classification with Many Labels. In Proceedings of the 1st GenBench Workshop on (Benchmarking) Generalisation in NLP, pages 173–184, Singapore. Association for Computational Linguistics.