@inproceedings{ren-etal-2026-personasteer,
title = "{P}ersona{S}teer: Dynamic Activation Steering for Personalized Multi-turn {LLM} Conversations",
author = "Ren, Sijian and
Zeng, Qingbin and
Gu, Bingqiao and
Xu, Fengli and
Li, Yong",
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.1270/",
doi = "10.18653/2026.findings-emnlp.1270",
pages = "26856--26876",
ISBN = "979-8-89176-462-0",
abstract = "Personalizing Large Language Models (LLMs) for long-term multi-turn interactions remains challenging, as user preferences and behavioral traits are progressively revealed and continuously evolving throughout conversations. Previous personalization methods have mostly relied on prompt-based or post-training (e.g., fine-tuning) paradigms; however, these approaches have been either prone to shallow alignment or have suffered from high computational costs. Inspired by the recent emergence of activation steering, we explore whether personalized alignment can be achieved through lightweight latent-space intervention without modifying backbone parameters. To achieve this goal, we propose \textbf{PersonaSteer}, a dynamic side-model architecture for multi-turn personalized alignment via learnable activation steering. Specifically, to enable fine-grained and evolving personalization, our design learns high-dimensional steering representations from unstructured interaction history and injects them into the backbone model{'}s residual stream for latent-space behavioral modulation. Furthermore, we introduce a steering-vector feedback loop to maintain persona consistency across multi-turn conversations. Extensive experiments demonstrate that PersonaSteer achieves best or top-tier alignment performance, averaging 3.18/5 on Qwen2.5-7B and 2.95/5 on Qwen2.5-3B. In addition, our framework approach maintains stability during evolving dialogues, with alignment performance consistently improving over time. PersonaSteer provides a lightweight solution for building truly dynamic and personalized LLM. The code is available in \url{https://github.com/tsinghua-fib-lab/PersonaSteer}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ren-etal-2026-personasteer">
<titleInfo>
<title>PersonaSteer: Dynamic Activation Steering for Personalized Multi-turn LLM Conversations</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sijian</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingbin</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bingqiao</namePart>
<namePart type="family">Gu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fengli</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yong</namePart>
<namePart type="family">Li</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>Personalizing Large Language Models (LLMs) for long-term multi-turn interactions remains challenging, as user preferences and behavioral traits are progressively revealed and continuously evolving throughout conversations. Previous personalization methods have mostly relied on prompt-based or post-training (e.g., fine-tuning) paradigms; however, these approaches have been either prone to shallow alignment or have suffered from high computational costs. Inspired by the recent emergence of activation steering, we explore whether personalized alignment can be achieved through lightweight latent-space intervention without modifying backbone parameters. To achieve this goal, we propose PersonaSteer, a dynamic side-model architecture for multi-turn personalized alignment via learnable activation steering. Specifically, to enable fine-grained and evolving personalization, our design learns high-dimensional steering representations from unstructured interaction history and injects them into the backbone model’s residual stream for latent-space behavioral modulation. Furthermore, we introduce a steering-vector feedback loop to maintain persona consistency across multi-turn conversations. Extensive experiments demonstrate that PersonaSteer achieves best or top-tier alignment performance, averaging 3.18/5 on Qwen2.5-7B and 2.95/5 on Qwen2.5-3B. In addition, our framework approach maintains stability during evolving dialogues, with alignment performance consistently improving over time. PersonaSteer provides a lightweight solution for building truly dynamic and personalized LLM. The code is available in https://github.com/tsinghua-fib-lab/PersonaSteer.</abstract>
<identifier type="citekey">ren-etal-2026-personasteer</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1270</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1270/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>26856</start>
<end>26876</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PersonaSteer: Dynamic Activation Steering for Personalized Multi-turn LLM Conversations
%A Ren, Sijian
%A Zeng, Qingbin
%A Gu, Bingqiao
%A Xu, Fengli
%A Li, Yong
%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 ren-etal-2026-personasteer
%X Personalizing Large Language Models (LLMs) for long-term multi-turn interactions remains challenging, as user preferences and behavioral traits are progressively revealed and continuously evolving throughout conversations. Previous personalization methods have mostly relied on prompt-based or post-training (e.g., fine-tuning) paradigms; however, these approaches have been either prone to shallow alignment or have suffered from high computational costs. Inspired by the recent emergence of activation steering, we explore whether personalized alignment can be achieved through lightweight latent-space intervention without modifying backbone parameters. To achieve this goal, we propose PersonaSteer, a dynamic side-model architecture for multi-turn personalized alignment via learnable activation steering. Specifically, to enable fine-grained and evolving personalization, our design learns high-dimensional steering representations from unstructured interaction history and injects them into the backbone model’s residual stream for latent-space behavioral modulation. Furthermore, we introduce a steering-vector feedback loop to maintain persona consistency across multi-turn conversations. Extensive experiments demonstrate that PersonaSteer achieves best or top-tier alignment performance, averaging 3.18/5 on Qwen2.5-7B and 2.95/5 on Qwen2.5-3B. In addition, our framework approach maintains stability during evolving dialogues, with alignment performance consistently improving over time. PersonaSteer provides a lightweight solution for building truly dynamic and personalized LLM. The code is available in https://github.com/tsinghua-fib-lab/PersonaSteer.
%R 10.18653/2026.findings-emnlp.1270
%U https://aclanthology.org/2026.findings-emnlp.1270/
%U https://doi.org/10.18653/2026.findings-emnlp.1270
%P 26856-26876
Markdown (Informal)
[PersonaSteer: Dynamic Activation Steering for Personalized Multi-turn LLM Conversations](https://aclanthology.org/2026.findings-emnlp.1270/) (Ren et al., Findings 2026)
ACL