@inproceedings{chen-etal-2026-asp,
title = "{ASP}: Adaptive Semantic Preservation for Context Compression",
author = "Chen, Shaoshen and
Li, Peiyang and
Mu, Jiaming and
Chen, Guixiong and
Li, Fengting and
Feng, Chao",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1084/",
pages = "22950--22967",
ISBN = "979-8-89176-462-0",
abstract = "Context compression has emerged as a vital approach to mitigate both performance degradation and resource consumption for large language models (LLMs) in long context scenarios. Current state-of-the-art methods allocate the compression budget at fixed-size block granularity, either uniformly or with adaptation across blocks. Both strategies fragment sentence-level semantic units and overlook the relevance variation within each block. In this paper, we propose the Adaptive Semantic Preservation (ASP) compressor, built on two innovations: (1) To preserve semantic units, our Semantic-Preserving Segmentation partitions context along natural sentence boundaries while maintaining the morphological atomicity of words. (2) To prioritize critical information, our task-aware Adaptive Budget Allocation dynamically assigns compression budget based on each sentence{'}s informational salience for the task instruction. Extensive experiments across diverse long-context scenarios show that ASP outperforms strong task-aware and task-agnostic compression baselines on both Llama-3-8B-Instruct and Qwen-2.5-7B-Instruct. In particular, ASP delivers stronger robustness compared with other baselines even at the extreme 16$\times$ high compression ratio."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-asp">
<titleInfo>
<title>ASP: Adaptive Semantic Preservation for Context Compression</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shaoshen</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peiyang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaming</namePart>
<namePart type="family">Mu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guixiong</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fengting</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chao</namePart>
<namePart type="family">Feng</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Context compression has emerged as a vital approach to mitigate both performance degradation and resource consumption for large language models (LLMs) in long context scenarios. Current state-of-the-art methods allocate the compression budget at fixed-size block granularity, either uniformly or with adaptation across blocks. Both strategies fragment sentence-level semantic units and overlook the relevance variation within each block. In this paper, we propose the Adaptive Semantic Preservation (ASP) compressor, built on two innovations: (1) To preserve semantic units, our Semantic-Preserving Segmentation partitions context along natural sentence boundaries while maintaining the morphological atomicity of words. (2) To prioritize critical information, our task-aware Adaptive Budget Allocation dynamically assigns compression budget based on each sentence’s informational salience for the task instruction. Extensive experiments across diverse long-context scenarios show that ASP outperforms strong task-aware and task-agnostic compression baselines on both Llama-3-8B-Instruct and Qwen-2.5-7B-Instruct. In particular, ASP delivers stronger robustness compared with other baselines even at the extreme 16\times high compression ratio.</abstract>
<identifier type="citekey">chen-etal-2026-asp</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1084/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>22950</start>
<end>22967</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T ASP: Adaptive Semantic Preservation for Context Compression
%A Chen, Shaoshen
%A Li, Peiyang
%A Mu, Jiaming
%A Chen, Guixiong
%A Li, Fengting
%A Feng, Chao
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F chen-etal-2026-asp
%X Context compression has emerged as a vital approach to mitigate both performance degradation and resource consumption for large language models (LLMs) in long context scenarios. Current state-of-the-art methods allocate the compression budget at fixed-size block granularity, either uniformly or with adaptation across blocks. Both strategies fragment sentence-level semantic units and overlook the relevance variation within each block. In this paper, we propose the Adaptive Semantic Preservation (ASP) compressor, built on two innovations: (1) To preserve semantic units, our Semantic-Preserving Segmentation partitions context along natural sentence boundaries while maintaining the morphological atomicity of words. (2) To prioritize critical information, our task-aware Adaptive Budget Allocation dynamically assigns compression budget based on each sentence’s informational salience for the task instruction. Extensive experiments across diverse long-context scenarios show that ASP outperforms strong task-aware and task-agnostic compression baselines on both Llama-3-8B-Instruct and Qwen-2.5-7B-Instruct. In particular, ASP delivers stronger robustness compared with other baselines even at the extreme 16\times high compression ratio.
%U https://aclanthology.org/2026.findings-emnlp.1084/
%P 22950-22967
Markdown (Informal)
[ASP: Adaptive Semantic Preservation for Context Compression](https://aclanthology.org/2026.findings-emnlp.1084/) (Chen et al., Findings 2026)
ACL
- Shaoshen Chen, Peiyang Li, Jiaming Mu, Guixiong Chen, Fengting Li, and Chao Feng. 2026. ASP: Adaptive Semantic Preservation for Context Compression. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 22950–22967, Budapest, Hungary. Association for Computational Linguistics.