@inproceedings{zhang-etal-2026-probe,
title = "{PROBE}: {PRO}cess-Based {BE}nchmark for Hallucination Detection",
author = "Zhang, Yu and
Belcak, Peter and
Diao, Shizhe and
Fu, Yonggan and
Ghosh, Shaona and
Mardani, Morteza and
Long, Eileen Margaret Peters and
Yu, Bei and
Molchanov, Pavlo",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.2099/",
pages = "42303--42320",
ISBN = "979-8-89176-395-1",
abstract = "Hallucination detection remains a significant challenge for large language models. Existing agentic applications rely on LLMs to self-assess the factuality of their outputs using single-step ``LLM-as-a-judge'' prompts. However, even when equipped with ground truth information, current LLMs still fall short in detecting hallucinations, and this one-shot evaluation offers neither the transparency nor the granularity needed to diagnose where and why the detection fails. To address this gap, we introduce PROBE (Process-based Benchmark for Hallucination Detection), a comprehensive benchmark that breaks down hallucination detection into four critical steps: claim decomposition, evidence finding, evidence evaluation, and hallucination localization, and evaluates each step individually. PROBE consists of 12,000 test cases across three task types{---}summarization, question answering, and style transfer. Critically, we demonstrate that when hallucination detection is treated as a multi-step process, all models achieve considerably better performance. Through extensive evaluation, we show that current LLMs struggle chiefly with evidence finding, and that finetuning on our released training data substantially improves performance on this step. PROBE represents a significant step toward more transparent, diagnosable, and robust hallucination detection systems."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-probe">
<titleInfo>
<title>PROBE: PROcess-Based BEnchmark for Hallucination Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Belcak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shizhe</namePart>
<namePart type="family">Diao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yonggan</namePart>
<namePart type="family">Fu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaona</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Morteza</namePart>
<namePart type="family">Mardani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eileen</namePart>
<namePart type="given">Margaret</namePart>
<namePart type="given">Peters</namePart>
<namePart type="family">Long</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bei</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pavlo</namePart>
<namePart type="family">Molchanov</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>Hallucination detection remains a significant challenge for large language models. Existing agentic applications rely on LLMs to self-assess the factuality of their outputs using single-step “LLM-as-a-judge” prompts. However, even when equipped with ground truth information, current LLMs still fall short in detecting hallucinations, and this one-shot evaluation offers neither the transparency nor the granularity needed to diagnose where and why the detection fails. To address this gap, we introduce PROBE (Process-based Benchmark for Hallucination Detection), a comprehensive benchmark that breaks down hallucination detection into four critical steps: claim decomposition, evidence finding, evidence evaluation, and hallucination localization, and evaluates each step individually. PROBE consists of 12,000 test cases across three task types—summarization, question answering, and style transfer. Critically, we demonstrate that when hallucination detection is treated as a multi-step process, all models achieve considerably better performance. Through extensive evaluation, we show that current LLMs struggle chiefly with evidence finding, and that finetuning on our released training data substantially improves performance on this step. PROBE represents a significant step toward more transparent, diagnosable, and robust hallucination detection systems.</abstract>
<identifier type="citekey">zhang-etal-2026-probe</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.2099/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>42303</start>
<end>42320</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PROBE: PROcess-Based BEnchmark for Hallucination Detection
%A Zhang, Yu
%A Belcak, Peter
%A Diao, Shizhe
%A Fu, Yonggan
%A Ghosh, Shaona
%A Mardani, Morteza
%A Long, Eileen Margaret Peters
%A Yu, Bei
%A Molchanov, Pavlo
%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 zhang-etal-2026-probe
%X Hallucination detection remains a significant challenge for large language models. Existing agentic applications rely on LLMs to self-assess the factuality of their outputs using single-step “LLM-as-a-judge” prompts. However, even when equipped with ground truth information, current LLMs still fall short in detecting hallucinations, and this one-shot evaluation offers neither the transparency nor the granularity needed to diagnose where and why the detection fails. To address this gap, we introduce PROBE (Process-based Benchmark for Hallucination Detection), a comprehensive benchmark that breaks down hallucination detection into four critical steps: claim decomposition, evidence finding, evidence evaluation, and hallucination localization, and evaluates each step individually. PROBE consists of 12,000 test cases across three task types—summarization, question answering, and style transfer. Critically, we demonstrate that when hallucination detection is treated as a multi-step process, all models achieve considerably better performance. Through extensive evaluation, we show that current LLMs struggle chiefly with evidence finding, and that finetuning on our released training data substantially improves performance on this step. PROBE represents a significant step toward more transparent, diagnosable, and robust hallucination detection systems.
%U https://aclanthology.org/2026.findings-acl.2099/
%P 42303-42320
Markdown (Informal)
[PROBE: PROcess-Based BEnchmark for Hallucination Detection](https://aclanthology.org/2026.findings-acl.2099/) (Zhang et al., Findings 2026)
ACL
- Yu Zhang, Peter Belcak, Shizhe Diao, Yonggan Fu, Shaona Ghosh, Morteza Mardani, Eileen Margaret Peters Long, Bei Yu, and Pavlo Molchanov. 2026. PROBE: PROcess-Based BEnchmark for Hallucination Detection. In Findings of the Association for Computational Linguistics: ACL 2026, pages 42303–42320, San Diego, California, United States. Association for Computational Linguistics.