Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs

Rareş-Alexandru Roşcan; Sergiu Nisioi

Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs

Correct Metadata for

Use this form to create a GitHub issue with structured data describing the correction. You will need a GitHub account. Once you create that issue, the correction will be reviewed by a staff member.

⚠️ Mobile Users: Submitting this form to create a new issue will only work with github.com, not the GitHub Mobile app.

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>. You may use <b>...</b> for bold, <i>...</i> for italic, and <url>...</url> for URLs.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

This paper presents our approach to the PARSEME 2.0 Shared Task on Romanian, covering both Identification (Subtask 1) and Paraphrasing (Subtask 2). While Large Language Models (LLMs) excel at semantic generation, we hypothesize that they lack the structural precision required for MWE identification, leading to "boundary hallucinations" that compromise downstream simplification. Our Rank 1 results on Romanian confirm this: a specialized encoder (RoBERT) using standard sequence labeling outperforms both few-shot LLMs and complex structural parsers (MTLB-STRUCT). This justifies our proposed pipeline: using encoders as precise “pointers” to guide the generative power of LLMs.

Anthology ID:: 2026.mwe-1.31
Volume:: Proceedings of the 22nd Workshop on Multiword Expressions (MWE 2026)
Month:: March
Year:: 2026
Address:: Rabat, Marocco
Editors:: Atul Kr. Ojha, Verginica Barbu Mititelu, Mathieu Constant, Ivelina Stoyanova, A. Seza Doğruöz, Alexandre Rademaker
Venues:: MWE | WS
SIG:
Publisher:: Association for Computational Linguistics
Note:
Pages:: 237–247
Language:
URL:: https://aclanthology.org/2026.mwe-1.31/
DOI:
Bibkey:
Cite (ACL):: Rares-Alexandru Roscan and Sergiu Nisioi. 2026. Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs. In Proceedings of the 22nd Workshop on Multiword Expressions (MWE 2026), pages 237–247, Rabat, Marocco. Association for Computational Linguistics.
Cite (Informal):: Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs (Roscan & Nisioi, MWE 2026)
Copy Citation:
PDF:: https://aclanthology.org/2026.mwe-1.31.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{roscan-nisioi-2026-archaeology,
    title = "Archaeology at {WE}-2026 {PARSEME} 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for {LLM}s",
    author = "Roscan, Rares-Alexandru  and
      Nisioi, Sergiu",
    editor = {Ojha, Atul Kr.  and
      Mititelu, Verginica Barbu  and
      Constant, Mathieu  and
      Stoyanova, Ivelina  and
      Do{\u{g}}ru{\"o}z, A. Seza  and
      Rademaker, Alexandre},
    booktitle = "Proceedings of the 22nd Workshop on Multiword Expressions ({MWE} 2026)",
    month = mar,
    year = "2026",
    address = "Rabat, Marocco",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.mwe-1.31/",
    pages = "237--247",
    ISBN = "979-8-89176-363-0",
    abstract = "This paper presents our approach to the PARSEME 2.0 Shared Task on Romanian, covering both Identification (Subtask 1) and Paraphrasing (Subtask 2). While Large Language Models (LLMs) excel at semantic generation, we hypothesize that they lack the structural precision required for MWE identification, leading to ``boundary hallucinations'' that compromise downstream simplification. Our Rank 1 results on Romanian confirm this: a specialized encoder (RoBERT) using standard sequence labeling outperforms both few-shot LLMs and complex structural parsers (MTLB-STRUCT). This justifies our proposed pipeline: using encoders as precise ``pointers'' to guide the generative power of LLMs."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="roscan-nisioi-2026-archaeology">
    <titleInfo>
        <title>Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Rares-Alexandru</namePart>
        <namePart type="family">Roscan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sergiu</namePart>
        <namePart type="family">Nisioi</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2026-03</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 22nd Workshop on Multiword Expressions (MWE 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Atul</namePart>
            <namePart type="given">Kr.</namePart>
            <namePart type="family">Ojha</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Verginica</namePart>
            <namePart type="given">Barbu</namePart>
            <namePart type="family">Mititelu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mathieu</namePart>
            <namePart type="family">Constant</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ivelina</namePart>
            <namePart type="family">Stoyanova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">A</namePart>
            <namePart type="given">Seza</namePart>
            <namePart type="family">Doğruöz</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Alexandre</namePart>
            <namePart type="family">Rademaker</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Rabat, Marocco</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-363-0</identifier>
    </relatedItem>
    <abstract>This paper presents our approach to the PARSEME 2.0 Shared Task on Romanian, covering both Identification (Subtask 1) and Paraphrasing (Subtask 2). While Large Language Models (LLMs) excel at semantic generation, we hypothesize that they lack the structural precision required for MWE identification, leading to “boundary hallucinations” that compromise downstream simplification. Our Rank 1 results on Romanian confirm this: a specialized encoder (RoBERT) using standard sequence labeling outperforms both few-shot LLMs and complex structural parsers (MTLB-STRUCT). This justifies our proposed pipeline: using encoders as precise “pointers” to guide the generative power of LLMs.</abstract>
    <identifier type="citekey">roscan-nisioi-2026-archaeology</identifier>
    <location>
        <url>https://aclanthology.org/2026.mwe-1.31/</url>
    </location>
    <part>
        <date>2026-03</date>
        <extent unit="page">
            <start>237</start>
            <end>247</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs
%A Roscan, Rares-Alexandru
%A Nisioi, Sergiu
%Y Ojha, Atul Kr.
%Y Mititelu, Verginica Barbu
%Y Constant, Mathieu
%Y Stoyanova, Ivelina
%Y Doğruöz, A. Seza
%Y Rademaker, Alexandre
%S Proceedings of the 22nd Workshop on Multiword Expressions (MWE 2026)
%D 2026
%8 March
%I Association for Computational Linguistics
%C Rabat, Marocco
%@ 979-8-89176-363-0
%F roscan-nisioi-2026-archaeology
%X This paper presents our approach to the PARSEME 2.0 Shared Task on Romanian, covering both Identification (Subtask 1) and Paraphrasing (Subtask 2). While Large Language Models (LLMs) excel at semantic generation, we hypothesize that they lack the structural precision required for MWE identification, leading to “boundary hallucinations” that compromise downstream simplification. Our Rank 1 results on Romanian confirm this: a specialized encoder (RoBERT) using standard sequence labeling outperforms both few-shot LLMs and complex structural parsers (MTLB-STRUCT). This justifies our proposed pipeline: using encoders as precise “pointers” to guide the generative power of LLMs.
%U https://aclanthology.org/2026.mwe-1.31/
%P 237-247

Download as File

Markdown (Informal)

[Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs](https://aclanthology.org/2026.mwe-1.31/) (Roscan & Nisioi, MWE 2026)

Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs (Roscan & Nisioi, MWE 2026)

ACL

Rares-Alexandru Roscan and Sergiu Nisioi. 2026. Archaeology at WE-2026 PARSEME 2.0 Subtask 1 and 2: Parsing is for Encoders, Paraphrasing is for LLMs. In Proceedings of the 22nd Workshop on Multiword Expressions (MWE 2026), pages 237–247, Rabat, Marocco. Association for Computational Linguistics.