<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chi-etal-2024-attention">
    <titleInfo>
        <title>Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Ta-Chung</namePart>
        <namePart type="family">Chi</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ting-Han</namePart>
        <namePart type="family">Fan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Alexander</namePart>
        <namePart type="family">Rudnicky</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2024-06</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: NAACL 2024</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Kevin</namePart>
            <namePart type="family">Duh</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Helena</namePart>
            <namePart type="family">Gomez</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Steven</namePart>
            <namePart type="family">Bethard</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Mexico City, Mexico</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>An ideal length-extrapolatable Transformer language model can handle sequences longer than the training length without any fine-tuning. Such long-context utilization capability relies heavily on a flexible positional embedding design. Upon investigating the flexibility of existing large pre-trained Transformer language models, we find that the T5 family deserves a closer look, as its positional embeddings capture rich and flexible attention patterns. However, T5 suffers from the dispersed attention issue: the longer the input sequence, the flatter the attention distribution. To alleviate the issue, we propose two attention alignment strategies via temperature scaling. Our findings show improvement on the long-context utilization capability of T5 on language modeling, retrieval, multi-document question answering, and code completion tasks without any fine-tuning. This suggests that a flexible positional embedding design and attention alignment can go a long way toward Transformer length extrapolation. The code is released at: https://github.com/chijames/T5-Attention-Alignment</abstract>
    <identifier type="citekey">chi-etal-2024-attention</identifier>
    <location>
        <url>https://aclanthology.org/2024.findings-naacl.10</url>
    </location>
    <part>
        <date>2024-06</date>
        <extent unit="page">
            <start>132</start>
            <end>148</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation
%A Chi, Ta-Chung
%A Fan, Ting-Han
%A Rudnicky, Alexander
%Y Duh, Kevin
%Y Gomez, Helena
%Y Bethard, Steven
%S Findings of the Association for Computational Linguistics: NAACL 2024
%D 2024
%8 June
%I Association for Computational Linguistics
%C Mexico City, Mexico
%F chi-etal-2024-attention
%X An ideal length-extrapolatable Transformer language model can handle sequences longer than the training length without any fine-tuning. Such long-context utilization capability relies heavily on a flexible positional embedding design. Upon investigating the flexibility of existing large pre-trained Transformer language models, we find that the T5 family deserves a closer look, as its positional embeddings capture rich and flexible attention patterns. However, T5 suffers from the dispersed attention issue: the longer the input sequence, the flatter the attention distribution. To alleviate the issue, we propose two attention alignment strategies via temperature scaling. Our findings show improvement on the long-context utilization capability of T5 on language modeling, retrieval, multi-document question answering, and code completion tasks without any fine-tuning. This suggests that a flexible positional embedding design and attention alignment can go a long way toward Transformer length extrapolation. The code is released at: https://github.com/chijames/T5-Attention-Alignment
%U https://aclanthology.org/2024.findings-naacl.10
%P 132-148

Markdown (Informal)

[Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation](https://aclanthology.org/2024.findings-naacl.10) (Chi et al., Findings 2024)

• Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation (Chi et al., Findings 2024)

ACL

• Ta-Chung Chi, Ting-Han Fan, and Alexander Rudnicky. 2024. Attention Alignment and Flexible Positional Embeddings Improve Transformer Length Extrapolation. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 132–148, Mexico City, Mexico. Association for Computational Linguistics.