<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shapiro-etal-2026-halt">
    <titleInfo>
        <title>HALT: Hallucination Assessment via Log-probs as Time series</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Ahmad</namePart>
        <namePart type="family">Shapiro</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Karan</namePart>
        <namePart type="family">Taneja</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ashok</namePart>
        <namePart type="family">Goel</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>Detecting hallucinations in Large Language Model (LLM) outputs is critical for safe deployment, yet existing methods require white-box access to internals, depend on multiple generations or auxiliary judge models, or collapse token-level uncertainty into scalar summaries that discard temporal structure. We introduce HALT, a lightweight grey-box detector that treats the top-20 log-probabilities of an LLM generation as a multivariate time series and processes them with a small bidirectional GRU (5M parameters), in a single forward pass and using no model internals. We evaluate on HUB, a unified protocol reorganizing existing hallucination datasets into ten capability clusters spanning factual and logical failure modes. HALT matches or surpasses Lettuce, a fine-tuned ModernBERT-base detector, on macro-F₁ while using \sim30\times fewer parameters; a controlled comparison against a text classifier trained on HALT’s own data confirms this advantage comes from the log-probability signal itself rather than from differences in training data, and the same recipe applies across eight LLM families from 360M to 70B parameters.</abstract>
    <identifier type="citekey">shapiro-etal-2026-halt</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.399/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>9174</start>
            <end>9190</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T HALT: Hallucination Assessment via Log-probs as Time series
%A Shapiro, Ahmad
%A Taneja, Karan
%A Goel, Ashok
%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 shapiro-etal-2026-halt
%X Detecting hallucinations in Large Language Model (LLM) outputs is critical for safe deployment, yet existing methods require white-box access to internals, depend on multiple generations or auxiliary judge models, or collapse token-level uncertainty into scalar summaries that discard temporal structure. We introduce HALT, a lightweight grey-box detector that treats the top-20 log-probabilities of an LLM generation as a multivariate time series and processes them with a small bidirectional GRU (5M parameters), in a single forward pass and using no model internals. We evaluate on HUB, a unified protocol reorganizing existing hallucination datasets into ten capability clusters spanning factual and logical failure modes. HALT matches or surpasses Lettuce, a fine-tuned ModernBERT-base detector, on macro-F₁ while using \sim30\times fewer parameters; a controlled comparison against a text classifier trained on HALT’s own data confirms this advantage comes from the log-probability signal itself rather than from differences in training data, and the same recipe applies across eight LLM families from 360M to 70B parameters.
%U https://aclanthology.org/2026.emnlp-main.399/
%P 9174-9190

Markdown (Informal)

[HALT: Hallucination Assessment via Log-probs as Time series](https://aclanthology.org/2026.emnlp-main.399/) (Shapiro et al., EMNLP 2026)

• HALT: Hallucination Assessment via Log-probs as Time series (Shapiro et al., EMNLP 2026)

ACL

• Ahmad Shapiro, Karan Taneja, and Ashok Goel. 2026. HALT: Hallucination Assessment via Log-probs as Time series. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 9174–9190, Budapest, Hungary. Association for Computational Linguistics.