@inproceedings{wang-zhao-2026-tare,
title = "{TARE}: Lightweight Token-Aware Representation Editing for Fine-tuning Transformer-like Models",
author = "Wang, Yulong and
Zhao, Siyu",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.841/",
doi = "10.18653/v1/2026.acl-long.841",
pages = "18452--18471",
ISBN = "979-8-89176-390-6",
abstract = "Parameter-efficient fine-tuning (PEFT) must balance effectiveness and efficiency: low-rank methods can be costly, while global representation edits often underfit token-level contexts. We propose **Token-Aware Representation Editing (TARE)**, a PEFT method that performs fine-grained, token-specific edits with a small additional inference overhead and minimal tuning.After each FFN block in a transformer-like model, we adopt a lightweight selector that scores a small pool of hidden representation editors for each token, activates only the top-k editors, and mixes their element-wise scaling/bias updates. This design achieves superior performance while maintaining computational efficiency, yielding a more favorable Pareto frontier compared to state-of-the-art (SOTA) methods.Across LLaMA-3-8B (eight knowledge reasoning and seven mathematical reasoning tasks) and RoBERTa-base/large (GLUE), TARE outperforms SOTAs (LoRA, DoRA, MiLoRA, LoReFT, and RED), achieving 86.7{\%} (knowledge reasoning), 76.7{\%} (mathematical reasoning), and 88.3{\%} (GLUE) while tuning only 0.0392{\%} of parameters using about 20 GiB peak GPU memory during training.An implementation is available at: {\ensuremath{<}}https://github.com/PatriciaPulec/tare{\ensuremath{>}}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-zhao-2026-tare">
<titleInfo>
<title>TARE: Lightweight Token-Aware Representation Editing for Fine-tuning Transformer-like Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yulong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siyu</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>Parameter-efficient fine-tuning (PEFT) must balance effectiveness and efficiency: low-rank methods can be costly, while global representation edits often underfit token-level contexts. We propose **Token-Aware Representation Editing (TARE)**, a PEFT method that performs fine-grained, token-specific edits with a small additional inference overhead and minimal tuning.After each FFN block in a transformer-like model, we adopt a lightweight selector that scores a small pool of hidden representation editors for each token, activates only the top-k editors, and mixes their element-wise scaling/bias updates. This design achieves superior performance while maintaining computational efficiency, yielding a more favorable Pareto frontier compared to state-of-the-art (SOTA) methods.Across LLaMA-3-8B (eight knowledge reasoning and seven mathematical reasoning tasks) and RoBERTa-base/large (GLUE), TARE outperforms SOTAs (LoRA, DoRA, MiLoRA, LoReFT, and RED), achieving 86.7% (knowledge reasoning), 76.7% (mathematical reasoning), and 88.3% (GLUE) while tuning only 0.0392% of parameters using about 20 GiB peak GPU memory during training.An implementation is available at: \ensuremath<https://github.com/PatriciaPulec/tare\ensuremath>.</abstract>
<identifier type="citekey">wang-zhao-2026-tare</identifier>
<identifier type="doi">10.18653/v1/2026.acl-long.841</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.841/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>18452</start>
<end>18471</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TARE: Lightweight Token-Aware Representation Editing for Fine-tuning Transformer-like Models
%A Wang, Yulong
%A Zhao, Siyu
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F wang-zhao-2026-tare
%X Parameter-efficient fine-tuning (PEFT) must balance effectiveness and efficiency: low-rank methods can be costly, while global representation edits often underfit token-level contexts. We propose **Token-Aware Representation Editing (TARE)**, a PEFT method that performs fine-grained, token-specific edits with a small additional inference overhead and minimal tuning.After each FFN block in a transformer-like model, we adopt a lightweight selector that scores a small pool of hidden representation editors for each token, activates only the top-k editors, and mixes their element-wise scaling/bias updates. This design achieves superior performance while maintaining computational efficiency, yielding a more favorable Pareto frontier compared to state-of-the-art (SOTA) methods.Across LLaMA-3-8B (eight knowledge reasoning and seven mathematical reasoning tasks) and RoBERTa-base/large (GLUE), TARE outperforms SOTAs (LoRA, DoRA, MiLoRA, LoReFT, and RED), achieving 86.7% (knowledge reasoning), 76.7% (mathematical reasoning), and 88.3% (GLUE) while tuning only 0.0392% of parameters using about 20 GiB peak GPU memory during training.An implementation is available at: \ensuremath<https://github.com/PatriciaPulec/tare\ensuremath>.
%R 10.18653/v1/2026.acl-long.841
%U https://aclanthology.org/2026.acl-long.841/
%U https://doi.org/10.18653/v1/2026.acl-long.841
%P 18452-18471
Markdown (Informal)
[TARE: Lightweight Token-Aware Representation Editing for Fine-tuning Transformer-like Models](https://aclanthology.org/2026.acl-long.841/) (Wang & Zhao, ACL 2026)
ACL