<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kim-etal-2026-libero">
    <titleInfo>
        <title>LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Chanyoung</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Minwoo</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Minseok</namePart>
        <namePart type="family">Kang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Hyunwoo</namePart>
        <namePart type="family">Kim</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Dahuin</namePart>
        <namePart type="family">Jung</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-10</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Hung-Yi</namePart>
            <namePart type="family">Lee</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sunipa</namePart>
            <namePart type="family">Dev</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Jessy</namePart>
            <namePart type="family">Li</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Desmond</namePart>
            <namePart type="family">Elliott</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Budapest, Hungary</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Vision–Language–Action (VLA) models achieve strong performance in robotic manipulation by leveraging pre-trained vision–language backbones. However, in downstream robotic settings, they are typically fine-tuned with limited data, leading to overfitting to specific instruction formulations and leaving robustness to paraphrased instructions underexplored. To study this gap, we introduce LIBERO-Para, a controlled benchmark that independently varies action expressions and object references for fine-grained analysis of linguistic generalization. Across seven VLA configurations (0.6B–7.5B), we observe consistent performance degradation of 22–52pp under paraphrasing. This degradation is primarily driven by object-level lexical variation: even simple synonym substitutions cause large drops, indicating reliance on surface-level matching rather than semantic grounding. Moreover, 80–96% of failures arise from planning-level trajectory divergence rather than execution errors, showing that paraphrasing disrupts task identification. Binary success rate treats all paraphrases equally, obscuring whether models perform consistently across difficulty levels or rely on easier cases. To address this, we propose PRIDE, a metric that quantifies paraphrase difficulty using semantic and syntactic factors. Our benchmark and code are available at https://github.com/cau-hai-lab/LIBERO-Para.</abstract>
    <identifier type="citekey">kim-etal-2026-libero</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.101/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>2378</start>
            <end>2414</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models
%A Kim, Chanyoung
%A Kim, Minwoo
%A Kang, Minseok
%A Kim, Hyunwoo
%A Jung, Dahuin
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F kim-etal-2026-libero
%X Vision–Language–Action (VLA) models achieve strong performance in robotic manipulation by leveraging pre-trained vision–language backbones. However, in downstream robotic settings, they are typically fine-tuned with limited data, leading to overfitting to specific instruction formulations and leaving robustness to paraphrased instructions underexplored. To study this gap, we introduce LIBERO-Para, a controlled benchmark that independently varies action expressions and object references for fine-grained analysis of linguistic generalization. Across seven VLA configurations (0.6B–7.5B), we observe consistent performance degradation of 22–52pp under paraphrasing. This degradation is primarily driven by object-level lexical variation: even simple synonym substitutions cause large drops, indicating reliance on surface-level matching rather than semantic grounding. Moreover, 80–96% of failures arise from planning-level trajectory divergence rather than execution errors, showing that paraphrasing disrupts task identification. Binary success rate treats all paraphrases equally, obscuring whether models perform consistently across difficulty levels or rely on easier cases. To address this, we propose PRIDE, a metric that quantifies paraphrase difficulty using semantic and syntactic factors. Our benchmark and code are available at https://github.com/cau-hai-lab/LIBERO-Para.
%U https://aclanthology.org/2026.emnlp-main.101/
%P 2378-2414

Markdown (Informal)

[LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models](https://aclanthology.org/2026.emnlp-main.101/) (Kim et al., EMNLP 2026)

• LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models (Kim et al., EMNLP 2026)

ACL

• Chanyoung Kim, Minwoo Kim, Minseok Kang, Hyunwoo Kim, and Dahuin Jung. 2026. LIBERO-Para: A Diagnostic Benchmark and Metrics for Paraphrase Robustness in VLA Models. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 2378–2414, Budapest, Hungary. Association for Computational Linguistics.