<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yamakoshi-etal-2023-causal">
    <titleInfo>
        <title>Causal interventions expose implicit situation models for commonsense language understanding</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Takateru</namePart>
        <namePart type="family">Yamakoshi</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">James</namePart>
        <namePart type="family">McClelland</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Adele</namePart>
        <namePart type="family">Goldberg</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Robert</namePart>
        <namePart type="family">Hawkins</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2023-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: ACL 2023</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Anna</namePart>
            <namePart type="family">Rogers</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jordan</namePart>
            <namePart type="family">Boyd-Graber</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Naoaki</namePart>
            <namePart type="family">Okazaki</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Toronto, Canada</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Accounts of human language processing have long appealed to implicit “situation models” that enrich comprehension with relevant but unstated world knowledge. Here, we apply causal intervention techniques to recent transformer models to analyze performance on the Winograd Schema Challenge (WSC), where a single context cue shifts interpretation of an ambiguous pronoun. We identify a relatively small circuit of attention heads that are responsible for propagating information from the context word that guides which of the candidate noun phrases the pronoun ultimately attends to. We then compare how this circuit behaves in a closely matched “syntactic” control where the situation model is not strictly necessary. These analyses suggest a distinct pathway through which implicit situation models may be constructed to guide pronoun resolution</abstract>
    <identifier type="citekey">yamakoshi-etal-2023-causal</identifier>
    <identifier type="doi">10.18653/v1/2023.findings-acl.839</identifier>
    <location>
        <url>https://aclanthology.org/2023.findings-acl.839</url>
    </location>
    <part>
        <date>2023-07</date>
        <extent unit="page">
            <start>13265</start>
            <end>13293</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Causal interventions expose implicit situation models for commonsense language understanding
%A Yamakoshi, Takateru
%A McClelland, James
%A Goldberg, Adele
%A Hawkins, Robert
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Findings of the Association for Computational Linguistics: ACL 2023
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F yamakoshi-etal-2023-causal
%X Accounts of human language processing have long appealed to implicit “situation models” that enrich comprehension with relevant but unstated world knowledge. Here, we apply causal intervention techniques to recent transformer models to analyze performance on the Winograd Schema Challenge (WSC), where a single context cue shifts interpretation of an ambiguous pronoun. We identify a relatively small circuit of attention heads that are responsible for propagating information from the context word that guides which of the candidate noun phrases the pronoun ultimately attends to. We then compare how this circuit behaves in a closely matched “syntactic” control where the situation model is not strictly necessary. These analyses suggest a distinct pathway through which implicit situation models may be constructed to guide pronoun resolution
%R 10.18653/v1/2023.findings-acl.839
%U https://aclanthology.org/2023.findings-acl.839
%U https://doi.org/10.18653/v1/2023.findings-acl.839
%P 13265-13293

Markdown (Informal)

[Causal interventions expose implicit situation models for commonsense language understanding](https://aclanthology.org/2023.findings-acl.839) (Yamakoshi et al., Findings 2023)

• Causal interventions expose implicit situation models for commonsense language understanding (Yamakoshi et al., Findings 2023)

ACL

• Takateru Yamakoshi, James McClelland, Adele Goldberg, and Robert Hawkins. 2023. Causal interventions expose implicit situation models for commonsense language understanding. In Findings of the Association for Computational Linguistics: ACL 2023, pages 13265–13293, Toronto, Canada. Association for Computational Linguistics.