@inproceedings{nakshatri-etal-2026-facts,
title = "When Facts Change: Probing {LLM}s on Evolving Knowledge with evolve{QA}",
author = "Nakshatri, Nishanth Sridhar and
Roy, Shamik and
Arivazhagan, Manoj Ghuhan and
Zhou, Hanhan and
Kumar, Vinayshekhar Bannihatti and
Gangadharaiah, Rashmi",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.2373/",
pages = "53864--53893",
ISBN = "979-8-89176-459-0",
abstract = "LLMs often fail to handle temporal knowledge conflicts{--}contradictions that arise when facts evolve over time within their training data. Existing benchmarks often rely on structured sources such as Wikidata, where changes are directly exposed through versioned snapshots, and rarely account for differences in model knowledge cut-offs, limiting fair evaluation across LLMs released at different times. We introduce a novel framework for automatically constructing temporal knowledge probing benchmarks from unstructured, time-stamped corpora by discovering evolving knowledge through salient entities, concepts, and attributes; and using this structure to generate grounded questions with gold answers aligned to model cut-off dates. Using this framework, we construct **evolveQA**, spanning 3 domains: AWS service updates, Azure platform changes, and WHO Disease Outbreak News reports. Evaluating 12 open- and closed-source LLMs across 3 probing formats, we find performance drops of up to 31{\%} on **evolveQA** compared to static knowledge questions, showing that LLMs struggle to recall temporally updated facts even when such knowledge is available in their parameters."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nakshatri-etal-2026-facts">
<titleInfo>
<title>When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nishanth</namePart>
<namePart type="given">Sridhar</namePart>
<namePart type="family">Nakshatri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shamik</namePart>
<namePart type="family">Roy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Manoj</namePart>
<namePart type="given">Ghuhan</namePart>
<namePart type="family">Arivazhagan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hanhan</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vinayshekhar</namePart>
<namePart type="given">Bannihatti</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rashmi</namePart>
<namePart type="family">Gangadharaiah</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>LLMs often fail to handle temporal knowledge conflicts–contradictions that arise when facts evolve over time within their training data. Existing benchmarks often rely on structured sources such as Wikidata, where changes are directly exposed through versioned snapshots, and rarely account for differences in model knowledge cut-offs, limiting fair evaluation across LLMs released at different times. We introduce a novel framework for automatically constructing temporal knowledge probing benchmarks from unstructured, time-stamped corpora by discovering evolving knowledge through salient entities, concepts, and attributes; and using this structure to generate grounded questions with gold answers aligned to model cut-off dates. Using this framework, we construct **evolveQA**, spanning 3 domains: AWS service updates, Azure platform changes, and WHO Disease Outbreak News reports. Evaluating 12 open- and closed-source LLMs across 3 probing formats, we find performance drops of up to 31% on **evolveQA** compared to static knowledge questions, showing that LLMs struggle to recall temporally updated facts even when such knowledge is available in their parameters.</abstract>
<identifier type="citekey">nakshatri-etal-2026-facts</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2373/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>53864</start>
<end>53893</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA
%A Nakshatri, Nishanth Sridhar
%A Roy, Shamik
%A Arivazhagan, Manoj Ghuhan
%A Zhou, Hanhan
%A Kumar, Vinayshekhar Bannihatti
%A Gangadharaiah, Rashmi
%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 nakshatri-etal-2026-facts
%X LLMs often fail to handle temporal knowledge conflicts–contradictions that arise when facts evolve over time within their training data. Existing benchmarks often rely on structured sources such as Wikidata, where changes are directly exposed through versioned snapshots, and rarely account for differences in model knowledge cut-offs, limiting fair evaluation across LLMs released at different times. We introduce a novel framework for automatically constructing temporal knowledge probing benchmarks from unstructured, time-stamped corpora by discovering evolving knowledge through salient entities, concepts, and attributes; and using this structure to generate grounded questions with gold answers aligned to model cut-off dates. Using this framework, we construct **evolveQA**, spanning 3 domains: AWS service updates, Azure platform changes, and WHO Disease Outbreak News reports. Evaluating 12 open- and closed-source LLMs across 3 probing formats, we find performance drops of up to 31% on **evolveQA** compared to static knowledge questions, showing that LLMs struggle to recall temporally updated facts even when such knowledge is available in their parameters.
%U https://aclanthology.org/2026.emnlp-main.2373/
%P 53864-53893
Markdown (Informal)
[When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA](https://aclanthology.org/2026.emnlp-main.2373/) (Nakshatri et al., EMNLP 2026)
ACL
- Nishanth Sridhar Nakshatri, Shamik Roy, Manoj Ghuhan Arivazhagan, Hanhan Zhou, Vinayshekhar Bannihatti Kumar, and Rashmi Gangadharaiah. 2026. When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 53864–53893, Budapest, Hungary. Association for Computational Linguistics.