<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ok-lee-2026-lost">
    <titleInfo>
        <title>Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Hyunjong</namePart>
        <namePart type="family">Ok</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Jaeho</namePart>
        <namePart type="family">Lee</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>Findings of the Association for Computational Linguistics: ACL 2026</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">979-8-89176-395-1</identifier>
    </relatedItem>
    <abstract>Large language models exhibit surprising sensitivity to the structure of the prompt, but the mechanisms underlying this sensitivity remain poorly understood. In this work, we conduct an in-depth investigation on a striking case: in multiple-choice question answering, placing context before the questions and options (CQO) outperforms the reverse order (QOC) by over 14%p, consistently over a wide range of models and datasets. Through systematic architectural analysis, we identify causal attention as the core mechanism: in QOC prompts, the causal mask prevents option tokens from attending to context, creating an information bottleneck where context becomes invisible to options.</abstract>
    <identifier type="citekey">ok-lee-2026-lost</identifier>
    <location>
        <url>https://aclanthology.org/2026.findings-acl.1921/</url>
    </location>
    <part>
        <date>2026-07</date>
        <extent unit="page">
            <start>38566</start>
            <end>38587</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models
%A Ok, Hyunjong
%A Lee, Jaeho
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F ok-lee-2026-lost
%X Large language models exhibit surprising sensitivity to the structure of the prompt, but the mechanisms underlying this sensitivity remain poorly understood. In this work, we conduct an in-depth investigation on a striking case: in multiple-choice question answering, placing context before the questions and options (CQO) outperforms the reverse order (QOC) by over 14%p, consistently over a wide range of models and datasets. Through systematic architectural analysis, we identify causal attention as the core mechanism: in QOC prompts, the causal mask prevents option tokens from attending to context, creating an information bottleneck where context becomes invisible to options.
%U https://aclanthology.org/2026.findings-acl.1921/
%P 38566-38587

Markdown (Informal)

[Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models](https://aclanthology.org/2026.findings-acl.1921/) (Ok & Lee, Findings 2026)

• Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models (Ok & Lee, Findings 2026)

ACL

• Hyunjong Ok and Jaeho Lee. 2026. Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models. In Findings of the Association for Computational Linguistics: ACL 2026, pages 38566–38587, San Diego, California, United States. Association for Computational Linguistics.