<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="adams-etal-2023-sparse">
    <titleInfo>
        <title>From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Griffin</namePart>
        <namePart type="family">Adams</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Alex</namePart>
        <namePart type="family">Fabbri</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Faisal</namePart>
        <namePart type="family">Ladhak</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Eric</namePart>
        <namePart type="family">Lehman</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Noémie</namePart>
        <namePart type="family">Elhadad</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 4th New Frontiers in Summarization Workshop</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Yue</namePart>
            <namePart type="family">Dong</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Wen</namePart>
            <namePart type="family">Xiao</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Lu</namePart>
            <namePart type="family">Wang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Fei</namePart>
            <namePart type="family">Liu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Giuseppe</namePart>
            <namePart type="family">Carenini</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>Selecting the “right” amount of information to include in a summary is a difficult task. A good summary should be detailed and entity-centric without being overly dense and hard to follow. To better understand this tradeoff, we solicit increasingly dense GPT-4 summaries with what we refer to as a “Chain of Density” (CoD) prompt. Specifically, GPT-4 generates an initial entity-sparse summary before iteratively incorporating missing salient entities without increasing the length. Summaries generated by CoD are more abstractive, exhibit more fusion, and have less of a lead bias than GPT-4 summaries generated by a vanilla prompt. We conduct a human preference study on 100 CNN DailyMail articles and find that humans prefer GPT-4 summaries that are more dense than those generated by a vanilla prompt and almost as dense as human written summaries. Qualitative analysis supports the notion that there exists a tradeoff between informativeness and readability. 500 annotated CoD summaries, as well as an extra 5,000 unannotated summaries, are freely available on HuggingFace (https://huggingface.co/datasets/griffin/chain_of_density).</abstract>
    <identifier type="citekey">adams-etal-2023-sparse</identifier>
    <identifier type="doi">10.18653/v1/2023.newsum-1.7</identifier>
    <location>
        <url>https://aclanthology.org/2023.newsum-1.7</url>
    </location>
    <part>
        <date>2023-12</date>
        <extent unit="page">
            <start>68</start>
            <end>74</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting
%A Adams, Griffin
%A Fabbri, Alex
%A Ladhak, Faisal
%A Lehman, Eric
%A Elhadad, Noémie
%Y Dong, Yue
%Y Xiao, Wen
%Y Wang, Lu
%Y Liu, Fei
%Y Carenini, Giuseppe
%S Proceedings of the 4th New Frontiers in Summarization Workshop
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F adams-etal-2023-sparse
%X Selecting the “right” amount of information to include in a summary is a difficult task. A good summary should be detailed and entity-centric without being overly dense and hard to follow. To better understand this tradeoff, we solicit increasingly dense GPT-4 summaries with what we refer to as a “Chain of Density” (CoD) prompt. Specifically, GPT-4 generates an initial entity-sparse summary before iteratively incorporating missing salient entities without increasing the length. Summaries generated by CoD are more abstractive, exhibit more fusion, and have less of a lead bias than GPT-4 summaries generated by a vanilla prompt. We conduct a human preference study on 100 CNN DailyMail articles and find that humans prefer GPT-4 summaries that are more dense than those generated by a vanilla prompt and almost as dense as human written summaries. Qualitative analysis supports the notion that there exists a tradeoff between informativeness and readability. 500 annotated CoD summaries, as well as an extra 5,000 unannotated summaries, are freely available on HuggingFace (https://huggingface.co/datasets/griffin/chain_of_density).
%R 10.18653/v1/2023.newsum-1.7
%U https://aclanthology.org/2023.newsum-1.7
%U https://doi.org/10.18653/v1/2023.newsum-1.7
%P 68-74

Markdown (Informal)

[From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting](https://aclanthology.org/2023.newsum-1.7) (Adams et al., NewSum 2023)

• From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting (Adams et al., NewSum 2023)

ACL

• Griffin Adams, Alex Fabbri, Faisal Ladhak, Eric Lehman, and Noémie Elhadad. 2023. From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting. In Proceedings of the 4th New Frontiers in Summarization Workshop, pages 68–74, Singapore. Association for Computational Linguistics.