@inproceedings{hu-etal-2025-stitchllm,
title = "{S}titch{LLM}: Serving {LLM}s, One Block at a Time",
author = "Hu, Bodun and
Li, Shuozhe and
Agarwal, Saurabh and
Lee, Myungjin and
Jajoo, Akshay and
Li, Jiamin and
Xu, Le and
Kim, Geon-Woo and
Kim, Donghyun and
Xu, Hong and
Zhang, Amy and
Akella, Aditya",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.1305/",
doi = "10.18653/v1/2025.acl-long.1305",
pages = "26887--26903",
ISBN = "979-8-89176-251-0",
abstract = "The rapid evolution of large language models (LLMs) has revolutionized natural language processing (NLP) tasks such as text generation, translation, and comprehension. However, the increasing computational demands and inference costs of these models present significant challenges. This study investigates the dynamic and efficient utilization of pre-trained weights from open-sourced LLMs of varying parameter sizes to achieve an optimal balance between computational efficiency and task performance. Drawing inspiration from the dual-process theory of human cognition, we introduce StitchLLM: a dynamic model routing framework that employs a powerful bottom model to process all queries, and uses a lightweight routing mechanism to allocate computational resources appropriately. Our novel framework optimizes efficiency and maintains performance, leveraging a trainable stitching layer for seamless integration of decoder layers across different LLMs. Experimental results demonstrate that StitchLLM improves system throughput while minimizing performance degradation, offering a flexible solution for deploying LLMs in resource-constrained settings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hu-etal-2025-stitchllm">
<titleInfo>
<title>StitchLLM: Serving LLMs, One Block at a Time</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bodun</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuozhe</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Saurabh</namePart>
<namePart type="family">Agarwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Myungjin</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Akshay</namePart>
<namePart type="family">Jajoo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiamin</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Le</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Geon-Woo</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Donghyun</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hong</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amy</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aditya</namePart>
<namePart type="family">Akella</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 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</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">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>The rapid evolution of large language models (LLMs) has revolutionized natural language processing (NLP) tasks such as text generation, translation, and comprehension. However, the increasing computational demands and inference costs of these models present significant challenges. This study investigates the dynamic and efficient utilization of pre-trained weights from open-sourced LLMs of varying parameter sizes to achieve an optimal balance between computational efficiency and task performance. Drawing inspiration from the dual-process theory of human cognition, we introduce StitchLLM: a dynamic model routing framework that employs a powerful bottom model to process all queries, and uses a lightweight routing mechanism to allocate computational resources appropriately. Our novel framework optimizes efficiency and maintains performance, leveraging a trainable stitching layer for seamless integration of decoder layers across different LLMs. Experimental results demonstrate that StitchLLM improves system throughput while minimizing performance degradation, offering a flexible solution for deploying LLMs in resource-constrained settings.</abstract>
<identifier type="citekey">hu-etal-2025-stitchllm</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.1305</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.1305/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>26887</start>
<end>26903</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T StitchLLM: Serving LLMs, One Block at a Time
%A Hu, Bodun
%A Li, Shuozhe
%A Agarwal, Saurabh
%A Lee, Myungjin
%A Jajoo, Akshay
%A Li, Jiamin
%A Xu, Le
%A Kim, Geon-Woo
%A Kim, Donghyun
%A Xu, Hong
%A Zhang, Amy
%A Akella, Aditya
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F hu-etal-2025-stitchllm
%X The rapid evolution of large language models (LLMs) has revolutionized natural language processing (NLP) tasks such as text generation, translation, and comprehension. However, the increasing computational demands and inference costs of these models present significant challenges. This study investigates the dynamic and efficient utilization of pre-trained weights from open-sourced LLMs of varying parameter sizes to achieve an optimal balance between computational efficiency and task performance. Drawing inspiration from the dual-process theory of human cognition, we introduce StitchLLM: a dynamic model routing framework that employs a powerful bottom model to process all queries, and uses a lightweight routing mechanism to allocate computational resources appropriately. Our novel framework optimizes efficiency and maintains performance, leveraging a trainable stitching layer for seamless integration of decoder layers across different LLMs. Experimental results demonstrate that StitchLLM improves system throughput while minimizing performance degradation, offering a flexible solution for deploying LLMs in resource-constrained settings.
%R 10.18653/v1/2025.acl-long.1305
%U https://aclanthology.org/2025.acl-long.1305/
%U https://doi.org/10.18653/v1/2025.acl-long.1305
%P 26887-26903
Markdown (Informal)
[StitchLLM: Serving LLMs, One Block at a Time](https://aclanthology.org/2025.acl-long.1305/) (Hu et al., ACL 2025)
ACL
- Bodun Hu, Shuozhe Li, Saurabh Agarwal, Myungjin Lee, Akshay Jajoo, Jiamin Li, Le Xu, Geon-Woo Kim, Donghyun Kim, Hong Xu, Amy Zhang, and Aditya Akella. 2025. StitchLLM: Serving LLMs, One Block at a Time. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 26887–26903, Vienna, Austria. Association for Computational Linguistics.