<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hendel-etal-2023-context">
    <titleInfo>
        <title>In-Context Learning Creates Task Vectors</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Roee</namePart>
        <namePart type="family">Hendel</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Mor</namePart>
        <namePart type="family">Geva</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Amir</namePart>
        <namePart type="family">Globerson</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>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Houda</namePart>
            <namePart type="family">Bouamor</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Juan</namePart>
            <namePart type="family">Pino</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Kalika</namePart>
            <namePart type="family">Bali</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) in Large Language Models (LLMs) has emerged as a powerful new learning paradigm. However, its underlying mechanism is still not well understood. In particular, it is challenging to map it to the “standard’ machine learning framework, where one uses a training set S to find a best-fitting function f(x) in some hypothesis class. Here we make progress on this problem by showing that the functions learned by ICL often have a very simple structure: they correspond to the transformer LLM whose only inputs are the query x and a single “task vector’ calculated from the training set. Thus, ICL can be seen as compressing S into a single task vector \boldsymbolθ(S) and then using this task vector to modulate the transformer to produce the output. We support the above claim via comprehensive experiments across a range of models and tasks.</abstract>
    <identifier type="citekey">hendel-etal-2023-context</identifier>
    <identifier type="doi">10.18653/v1/2023.findings-emnlp.624</identifier>
    <location>
        <url>https://aclanthology.org/2023.findings-emnlp.624</url>
    </location>
    <part>
        <date>2023-12</date>
        <extent unit="page">
            <start>9318</start>
            <end>9333</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T In-Context Learning Creates Task Vectors
%A Hendel, Roee
%A Geva, Mor
%A Globerson, Amir
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F hendel-etal-2023-context
%X In-context learning (ICL) in Large Language Models (LLMs) has emerged as a powerful new learning paradigm. However, its underlying mechanism is still not well understood. In particular, it is challenging to map it to the “standard’ machine learning framework, where one uses a training set S to find a best-fitting function f(x) in some hypothesis class. Here we make progress on this problem by showing that the functions learned by ICL often have a very simple structure: they correspond to the transformer LLM whose only inputs are the query x and a single “task vector’ calculated from the training set. Thus, ICL can be seen as compressing S into a single task vector \boldsymbolθ(S) and then using this task vector to modulate the transformer to produce the output. We support the above claim via comprehensive experiments across a range of models and tasks.
%R 10.18653/v1/2023.findings-emnlp.624
%U https://aclanthology.org/2023.findings-emnlp.624
%U https://doi.org/10.18653/v1/2023.findings-emnlp.624
%P 9318-9333

Markdown (Informal)

[In-Context Learning Creates Task Vectors](https://aclanthology.org/2023.findings-emnlp.624) (Hendel et al., Findings 2023)

• In-Context Learning Creates Task Vectors (Hendel et al., Findings 2023)

ACL

• Roee Hendel, Mor Geva, and Amir Globerson. 2023. In-Context Learning Creates Task Vectors. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 9318–9333, Singapore. Association for Computational Linguistics.