<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rozanova-etal-2023-interventional">
    <titleInfo>
        <title>Interventional Probing in High Dimensions: An NLI Case Study</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Julia</namePart>
        <namePart type="family">Rozanova</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Marco</namePart>
        <namePart type="family">Valentino</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Lucas</namePart>
        <namePart type="family">Cordeiro</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">André</namePart>
        <namePart type="family">Freitas</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2023-05</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Findings of the Association for Computational Linguistics: EACL 2023</title>
        </titleInfo>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Dubrovnik, Croatia</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>Probing strategies have been shown to detectthe presence of various linguistic features inlarge language models; in particular, seman-tic features intermediate to the “natural logic”fragment of the Natural Language Inferencetask (NLI). In the case of natural logic, the rela-tion between the intermediate features and theentailment label is explicitly known: as such,this provides a ripe setting for interventionalstudies on the NLI models’ representations, al-lowing for stronger causal conjectures and adeeper critical analysis of interventional prob-ing methods. In this work, we carry out newand existing representation-level interventionsto investigate the effect of these semantic fea-tures on NLI classification: we perform am-nesic probing (which removes features as di-rected by learned linear probes) and introducethe mnestic probing variation (which forgetsall dimensions except the probe-selected ones).Furthermore, we delve into the limitations ofthese methods and outline some pitfalls havebeen obscuring the effectivity of interventionalprobing studies.</abstract>
    <identifier type="citekey">rozanova-etal-2023-interventional</identifier>
    <identifier type="doi">10.18653/v1/2023.findings-eacl.188</identifier>
    <location>
        <url>https://aclanthology.org/2023.findings-eacl.188</url>
    </location>
    <part>
        <date>2023-05</date>
        <extent unit="page">
            <start>2489</start>
            <end>2500</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Interventional Probing in High Dimensions: An NLI Case Study
%A Rozanova, Julia
%A Valentino, Marco
%A Cordeiro, Lucas
%A Freitas, André
%S Findings of the Association for Computational Linguistics: EACL 2023
%D 2023
%8 May
%I Association for Computational Linguistics
%C Dubrovnik, Croatia
%F rozanova-etal-2023-interventional
%X Probing strategies have been shown to detectthe presence of various linguistic features inlarge language models; in particular, seman-tic features intermediate to the “natural logic”fragment of the Natural Language Inferencetask (NLI). In the case of natural logic, the rela-tion between the intermediate features and theentailment label is explicitly known: as such,this provides a ripe setting for interventionalstudies on the NLI models’ representations, al-lowing for stronger causal conjectures and adeeper critical analysis of interventional prob-ing methods. In this work, we carry out newand existing representation-level interventionsto investigate the effect of these semantic fea-tures on NLI classification: we perform am-nesic probing (which removes features as di-rected by learned linear probes) and introducethe mnestic probing variation (which forgetsall dimensions except the probe-selected ones).Furthermore, we delve into the limitations ofthese methods and outline some pitfalls havebeen obscuring the effectivity of interventionalprobing studies.
%R 10.18653/v1/2023.findings-eacl.188
%U https://aclanthology.org/2023.findings-eacl.188
%U https://doi.org/10.18653/v1/2023.findings-eacl.188
%P 2489-2500

Markdown (Informal)

[Interventional Probing in High Dimensions: An NLI Case Study](https://aclanthology.org/2023.findings-eacl.188) (Rozanova et al., Findings 2023)

• Interventional Probing in High Dimensions: An NLI Case Study (Rozanova et al., Findings 2023)

ACL

• Julia Rozanova, Marco Valentino, Lucas Cordeiro, and André Freitas. 2023. Interventional Probing in High Dimensions: An NLI Case Study. In Findings of the Association for Computational Linguistics: EACL 2023, pages 2489–2500, Dubrovnik, Croatia. Association for Computational Linguistics.