<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="cui-etal-2026-llms">
    <titleInfo>
        <title>LLMs Regret Before They Say It: Early Detection and Compositional Architecture of Regret in Hidden States</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Xiangxiang</namePart>
        <namePart type="family">Cui</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shu</namePart>
        <namePart type="family">Yang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Tianjin</namePart>
        <namePart type="family">Huang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Wanyu</namePart>
        <namePart type="family">Lin</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Lijie</namePart>
        <namePart type="family">Hu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Di</namePart>
        <namePart type="family">Wang</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
    </relatedItem>
    <abstract>Regret in large language models (LLMs) refers to their explicit expression of regret when confronted with evidence contradicting previously generated misinformation. Understanding the neural encoding of regret is crucial for advancing artificial metacognition and model reliability. However, three key gaps remain: the absence of specialized datasets for regret expressions, the lack of understanding of how early regret signals emerge in hidden states, and the unknown neuron-level architecture underlying regret encoding. To address these gaps, we make three contributions: (1) we propose a multi-stage workflow for constructing the first regret dataset and train probes achieving near-perfect accuracy across LLaMA-2-7B, Mistral-7B, and Vicuna-13B. (2) we discover that regret signals begin emerging in preceding hidden states before the regret token appears, and that probes generalize to semantically related keywords such as “incorrect” and “apologize.” (3) at the neuron level, we introduce the Regret Dominance Score (RDS) and Group Impact Coefficient (GIC) to reveal a compositional regret architecture where cooperative neuron groups collectively encode regret, with larger models (70B) exhibiting more pronounced compositional effects. Targeted interventions on these neurons successfully suppress regret expression in generated text, validating our mechanistic understanding. Our code and data are available at https://github.com/AbnerAI/LLMs-Regret-Before-They-Say-It.</abstract>
    <identifier type="citekey">cui-etal-2026-llms</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.215/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>4923</start>
            <end>4946</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T LLMs Regret Before They Say It: Early Detection and Compositional Architecture of Regret in Hidden States
%A Cui, Xiangxiang
%A Yang, Shu
%A Huang, Tianjin
%A Lin, Wanyu
%A Hu, Lijie
%A Wang, Di
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F cui-etal-2026-llms
%X Regret in large language models (LLMs) refers to their explicit expression of regret when confronted with evidence contradicting previously generated misinformation. Understanding the neural encoding of regret is crucial for advancing artificial metacognition and model reliability. However, three key gaps remain: the absence of specialized datasets for regret expressions, the lack of understanding of how early regret signals emerge in hidden states, and the unknown neuron-level architecture underlying regret encoding. To address these gaps, we make three contributions: (1) we propose a multi-stage workflow for constructing the first regret dataset and train probes achieving near-perfect accuracy across LLaMA-2-7B, Mistral-7B, and Vicuna-13B. (2) we discover that regret signals begin emerging in preceding hidden states before the regret token appears, and that probes generalize to semantically related keywords such as “incorrect” and “apologize.” (3) at the neuron level, we introduce the Regret Dominance Score (RDS) and Group Impact Coefficient (GIC) to reveal a compositional regret architecture where cooperative neuron groups collectively encode regret, with larger models (70B) exhibiting more pronounced compositional effects. Targeted interventions on these neurons successfully suppress regret expression in generated text, validating our mechanistic understanding. Our code and data are available at https://github.com/AbnerAI/LLMs-Regret-Before-They-Say-It.
%U https://aclanthology.org/2026.emnlp-main.215/
%P 4923-4946

Markdown (Informal)

[LLMs Regret Before They Say It: Early Detection and Compositional Architecture of Regret in Hidden States](https://aclanthology.org/2026.emnlp-main.215/) (Cui et al., EMNLP 2026)

• LLMs Regret Before They Say It: Early Detection and Compositional Architecture of Regret in Hidden States (Cui et al., EMNLP 2026)

ACL

• Xiangxiang Cui, Shu Yang, Tianjin Huang, Wanyu Lin, Lijie Hu, and Di Wang. 2026. LLMs Regret Before They Say It: Early Detection and Compositional Architecture of Regret in Hidden States. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 4923–4946, Budapest, Hungary. Association for Computational Linguistics.