<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="qiu-jialingan-2026-tokendecouple">
    <titleInfo>
        <title>TokenDecouple: Frequency-aware Token Merging and Pyramidal Text Decay</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Xin</namePart>
        <namePart type="family">Qiu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name>
        <namePart>Jialingan</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>Large Language Models (LLMs) have recently shown strong potential for time series (TS) analysis by integrating numerical signals with rich text prompts. In this process, although TS are encoded into token sequences and fed into LLMs in a manner similar to text prompts, TS tokens differ from prompt tokens in their underlying structure and semantics, a distinction that remains underexplored in existing LLM-based TS systems. In this work, we investigate both TS tokens and prompt tokens from a new perspective. First, we re-examine TS tokens in the frequency-domain and reveal that token-level information contribution is highly uneven: large groups of tokens share redundant spectral patterns, while a small subset acts as critical frequency anchors. Motivated by this observation, we introduce a frequency-based token merging strategy for TS tokens, which preserves essential information while eliminating redundancy. Second, we uncover a unexplored pyramidal decay effect in prompt tokens, showing that their influence rapidly diminishes with model depth. Building on these findings, we propose TokenDecouple, a unified framework that decouples the compression of TS tokens and prompt tokens. Extensive experiments across diverse TS tasks show that TokenDecouple achieves up to 7.68× inference speedup while delivering performance improvements in 78% of evaluated scenarios, establishing an efficient paradigm for LLM-based TS analysis.</abstract>
    <identifier type="citekey">qiu-jialingan-2026-tokendecouple</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.1787/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>40718</start>
            <end>40741</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T TokenDecouple: Frequency-aware Token Merging and Pyramidal Text Decay
%A Qiu, Xin
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%A Jialingan
%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 qiu-jialingan-2026-tokendecouple
%X Large Language Models (LLMs) have recently shown strong potential for time series (TS) analysis by integrating numerical signals with rich text prompts. In this process, although TS are encoded into token sequences and fed into LLMs in a manner similar to text prompts, TS tokens differ from prompt tokens in their underlying structure and semantics, a distinction that remains underexplored in existing LLM-based TS systems. In this work, we investigate both TS tokens and prompt tokens from a new perspective. First, we re-examine TS tokens in the frequency-domain and reveal that token-level information contribution is highly uneven: large groups of tokens share redundant spectral patterns, while a small subset acts as critical frequency anchors. Motivated by this observation, we introduce a frequency-based token merging strategy for TS tokens, which preserves essential information while eliminating redundancy. Second, we uncover a unexplored pyramidal decay effect in prompt tokens, showing that their influence rapidly diminishes with model depth. Building on these findings, we propose TokenDecouple, a unified framework that decouples the compression of TS tokens and prompt tokens. Extensive experiments across diverse TS tasks show that TokenDecouple achieves up to 7.68× inference speedup while delivering performance improvements in 78% of evaluated scenarios, establishing an efficient paradigm for LLM-based TS analysis.
%U https://aclanthology.org/2026.emnlp-main.1787/
%P 40718-40741

Markdown (Informal)

[TokenDecouple: Frequency-aware Token Merging and Pyramidal Text Decay](https://aclanthology.org/2026.emnlp-main.1787/) (Qiu & Jialingan, EMNLP 2026)

• TokenDecouple: Frequency-aware Token Merging and Pyramidal Text Decay (Qiu & Jialingan, EMNLP 2026)

ACL

• Xin Qiu and Jialingan. 2026. TokenDecouple: Frequency-aware Token Merging and Pyramidal Text Decay. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 40718–40741, Budapest, Hungary. Association for Computational Linguistics.