TableKV: KV Cache Compression for In-Context Table Processing

Giulio Corallo; Elia Faure-Rolland; Miriam Lamari; Paolo Papotti

doi:10.18653/v1/2025.trl-1.13

TableKV: KV Cache Compression for In-Context Table Processing

Giulio Corallo, Elia Faure-Rolland, Miriam Lamari, Paolo Papotti

Correct Metadata for

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>.

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

Processing large tables provided in-context to LLMs is challenging due to token limits and information overload. While Retrieval-Augmented Generation can select relevant subsets externally, this work explores Key-Value (KV) cache compression as an alternative, applied directly to the linearized table during inference. We show that the LLM’s internal attention scores over the table context guides the retention of essential KV pairs, effectively compressing the processing context while preserving crucial relational information needed for complex queries. Experiments on Spider, WikitableQA, and QTSumm datasets validate the compression approach for in-context table processing, offering a promising path for improved table representation learning in LLMs.

Anthology ID:: 2025.trl-1.13
Volume:: Proceedings of the 4th Table Representation Learning Workshop
Month:: July
Year:: 2025
Address:: Vienna, Austria
Editors:: Shuaichen Chang, Madelon Hulsebos, Qian Liu, Wenhu Chen, Huan Sun
Venues:: TRL | WS
SIG:
Publisher:: Association for Computational Linguistics
Note:
Pages:: 166–171
Language:
URL:: https://aclanthology.org/2025.trl-1.13/
DOI:: 10.18653/v1/2025.trl-1.13
Bibkey:
Cite (ACL):: Giulio Corallo, Elia Faure-Rolland, Miriam Lamari, and Paolo Papotti. 2025. TableKV: KV Cache Compression for In-Context Table Processing. In Proceedings of the 4th Table Representation Learning Workshop, pages 166–171, Vienna, Austria. Association for Computational Linguistics.
Cite (Informal):: TableKV: KV Cache Compression for In-Context Table Processing (Corallo et al., TRL 2025)
Copy Citation:
PDF:: https://aclanthology.org/2025.trl-1.13.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{corallo-etal-2025-tablekv,
    title = "{T}able{KV}: {KV} Cache Compression for In-Context Table Processing",
    author = "Corallo, Giulio  and
      Faure-Rolland, Elia  and
      Lamari, Miriam  and
      Papotti, Paolo",
    editor = "Chang, Shuaichen  and
      Hulsebos, Madelon  and
      Liu, Qian  and
      Chen, Wenhu  and
      Sun, Huan",
    booktitle = "Proceedings of the 4th Table Representation Learning Workshop",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.trl-1.13/",
    doi = "10.18653/v1/2025.trl-1.13",
    pages = "166--171",
    ISBN = "979-8-89176-268-8",
    abstract = "Processing large tables provided in-context to LLMs is challenging due to token limits and information overload. While Retrieval-Augmented Generation can select relevant subsets externally, this work explores Key-Value (KV) cache compression as an alternative, applied directly to the linearized table during inference. We show that the LLM{'}s internal attention scores over the table context guides the retention of essential KV pairs, effectively compressing the processing context while preserving crucial relational information needed for complex queries. Experiments on Spider, WikitableQA, and QTSumm datasets validate the compression approach for in-context table processing, offering a promising path for improved table representation learning in LLMs."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="corallo-etal-2025-tablekv">
    <titleInfo>
        <title>TableKV: KV Cache Compression for In-Context Table Processing</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Giulio</namePart>
        <namePart type="family">Corallo</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Elia</namePart>
        <namePart type="family">Faure-Rolland</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Miriam</namePart>
        <namePart type="family">Lamari</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Paolo</namePart>
        <namePart type="family">Papotti</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 4th Table Representation Learning Workshop</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Shuaichen</namePart>
            <namePart type="family">Chang</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Madelon</namePart>
            <namePart type="family">Hulsebos</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Qian</namePart>
            <namePart type="family">Liu</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Wenhu</namePart>
            <namePart type="family">Chen</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Huan</namePart>
            <namePart type="family">Sun</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Vienna, Austria</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-268-8</identifier>
    </relatedItem>
    <abstract>Processing large tables provided in-context to LLMs is challenging due to token limits and information overload. While Retrieval-Augmented Generation can select relevant subsets externally, this work explores Key-Value (KV) cache compression as an alternative, applied directly to the linearized table during inference. We show that the LLM’s internal attention scores over the table context guides the retention of essential KV pairs, effectively compressing the processing context while preserving crucial relational information needed for complex queries. Experiments on Spider, WikitableQA, and QTSumm datasets validate the compression approach for in-context table processing, offering a promising path for improved table representation learning in LLMs.</abstract>
    <identifier type="citekey">corallo-etal-2025-tablekv</identifier>
    <identifier type="doi">10.18653/v1/2025.trl-1.13</identifier>
    <location>
        <url>https://aclanthology.org/2025.trl-1.13/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>166</start>
            <end>171</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T TableKV: KV Cache Compression for In-Context Table Processing
%A Corallo, Giulio
%A Faure-Rolland, Elia
%A Lamari, Miriam
%A Papotti, Paolo
%Y Chang, Shuaichen
%Y Hulsebos, Madelon
%Y Liu, Qian
%Y Chen, Wenhu
%Y Sun, Huan
%S Proceedings of the 4th Table Representation Learning Workshop
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-268-8
%F corallo-etal-2025-tablekv
%X Processing large tables provided in-context to LLMs is challenging due to token limits and information overload. While Retrieval-Augmented Generation can select relevant subsets externally, this work explores Key-Value (KV) cache compression as an alternative, applied directly to the linearized table during inference. We show that the LLM’s internal attention scores over the table context guides the retention of essential KV pairs, effectively compressing the processing context while preserving crucial relational information needed for complex queries. Experiments on Spider, WikitableQA, and QTSumm datasets validate the compression approach for in-context table processing, offering a promising path for improved table representation learning in LLMs.
%R 10.18653/v1/2025.trl-1.13
%U https://aclanthology.org/2025.trl-1.13/
%U https://doi.org/10.18653/v1/2025.trl-1.13
%P 166-171

Download as File

Markdown (Informal)

[TableKV: KV Cache Compression for In-Context Table Processing](https://aclanthology.org/2025.trl-1.13/) (Corallo et al., TRL 2025)

TableKV: KV Cache Compression for In-Context Table Processing (Corallo et al., TRL 2025)

ACL

Giulio Corallo, Elia Faure-Rolland, Miriam Lamari, and Paolo Papotti. 2025. TableKV: KV Cache Compression for In-Context Table Processing. In Proceedings of the 4th Table Representation Learning Workshop, pages 166–171, Vienna, Austria. Association for Computational Linguistics.