<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="steele-2026-annotation">
    <titleInfo>
        <title>Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Brady</namePart>
        <namePart type="family">Steele</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 (ACL 2026)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Santosh</namePart>
            <namePart type="family">T.Y.S.S.</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Juan</namePart>
            <namePart type="given">Diego</namePart>
            <namePart type="family">Rodriguez</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ona</namePart>
            <namePart type="family">de Gibert</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">979-8-89176-393-7</identifier>
    </relatedItem>
    <abstract>Annotator disagreement on tasks like natural languageinference (NLI) reflects genuine linguistic ambiguity,yet most fine-tuning recipes treat every example as equallylearnable.We ask whether this external signal of ambiguity predicts*per-example* learning dynamics under LoRA, the most widelyused parameter-efficient fine-tuning method, and find that it does.Correlating annotation entropy (from ChaosNLI’s 100 labels perexample) with per-example area under the loss curve (AULC)on SNLI and MNLI, the correlation is positive in all 25conditions tested (Spearman ρ = 0.06-0.43), withdecoder-only models showing stronger correlations thanencoders at matched LoRA rank.More strikingly, under LoRA contested examples exhibit*un-learning*: their gold-label loss *increases*during training, a pattern that is largely absent underfull fine-tuning and IA³ in the encoder setting wherematched comparisons are available, and that we also observeunder LoRA on two decoder-only models.The effect survives partial-correlation controls andreplicates across seeds and datasets.A preliminary noise-injection experiment is consistentwith these findings.</abstract>
    <identifier type="citekey">steele-2026-annotation</identifier>
    <location>
        <url>https://aclanthology.org/2026.acl-srw.11/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>129</start>
            <end>141</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning
%A Steele, Brady
%Y T.Y.S.S., Santosh
%Y Rodriguez, Juan Diego
%Y de Gibert, Ona
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-393-7
%F steele-2026-annotation
%X Annotator disagreement on tasks like natural languageinference (NLI) reflects genuine linguistic ambiguity,yet most fine-tuning recipes treat every example as equallylearnable.We ask whether this external signal of ambiguity predicts*per-example* learning dynamics under LoRA, the most widelyused parameter-efficient fine-tuning method, and find that it does.Correlating annotation entropy (from ChaosNLI’s 100 labels perexample) with per-example area under the loss curve (AULC)on SNLI and MNLI, the correlation is positive in all 25conditions tested (Spearman ρ = 0.06-0.43), withdecoder-only models showing stronger correlations thanencoders at matched LoRA rank.More strikingly, under LoRA contested examples exhibit*un-learning*: their gold-label loss *increases*during training, a pattern that is largely absent underfull fine-tuning and IA³ in the encoder setting wherematched comparisons are available, and that we also observeunder LoRA on two decoder-only models.The effect survives partial-correlation controls andreplicates across seeds and datasets.A preliminary noise-injection experiment is consistentwith these findings.
%U https://aclanthology.org/2026.acl-srw.11/
%P 129-141

Markdown (Informal)

[Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning](https://aclanthology.org/2026.acl-srw.11/) (Steele, ACL 2026)

• Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning (Steele, ACL 2026)

ACL

• Brady Steele. 2026. Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), pages 129–141, San Diego, California, United States. Association for Computational Linguistics.