@inproceedings{rahimi-etal-2026-reframe,
title = "{REFRAME}: A Test Suite for the Knowledge{--}Deployment Gap in {LLM}s",
author = "Rahimi, Elahe and
Roewer-Despr{\'e}s, Fran{\c{c}}ois and
Sajjad, Hassan and
Alam, Md Mahbub and
Badshah, Sher and
Dolatabadi, Elham and
Rudzicz, Frank",
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.2520/",
pages = "57310--57325",
ISBN = "979-8-89176-459-0",
abstract = "Large language models (LLMs) are commonly evaluated on multiple-choice tasks that ask them to select an answer or to abstain on unanswerable questions. Neither setting captures frame-breaking: where the question is answerable, but the correct answer lies outside the framing of a candidate set of invalid answers. A useful model must reframe the task, rather than choose an invalid frame-compliant answer. We introduce REFRAME, a test suite of 150 scenarios in which a commonsense constraint invalidates their framing. Each REFRAME scenario is additionally paired with a direct knowledge probe of the same frame-breaking constraint, which we leverage to separate what a model knows from what it deploys. Across nine frontier LLMs, we observe a substantial knowledge-deployment gap with zero-shot prompting. A targeted reasoning prompt only partially closes this gap. Human annotators, by contrast, readily detect the constraint and reframe the problem. Together, these results characterize frame-breaking as a distinct reasoning challenge and demonstrate significant room for LLM improvement."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rahimi-etal-2026-reframe">
<titleInfo>
<title>REFRAME: A Test Suite for the Knowledge–Deployment Gap in LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Elahe</namePart>
<namePart type="family">Rahimi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">François</namePart>
<namePart type="family">Roewer-Després</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hassan</namePart>
<namePart type="family">Sajjad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Md</namePart>
<namePart type="given">Mahbub</namePart>
<namePart type="family">Alam</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sher</namePart>
<namePart type="family">Badshah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elham</namePart>
<namePart type="family">Dolatabadi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Frank</namePart>
<namePart type="family">Rudzicz</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>Large language models (LLMs) are commonly evaluated on multiple-choice tasks that ask them to select an answer or to abstain on unanswerable questions. Neither setting captures frame-breaking: where the question is answerable, but the correct answer lies outside the framing of a candidate set of invalid answers. A useful model must reframe the task, rather than choose an invalid frame-compliant answer. We introduce REFRAME, a test suite of 150 scenarios in which a commonsense constraint invalidates their framing. Each REFRAME scenario is additionally paired with a direct knowledge probe of the same frame-breaking constraint, which we leverage to separate what a model knows from what it deploys. Across nine frontier LLMs, we observe a substantial knowledge-deployment gap with zero-shot prompting. A targeted reasoning prompt only partially closes this gap. Human annotators, by contrast, readily detect the constraint and reframe the problem. Together, these results characterize frame-breaking as a distinct reasoning challenge and demonstrate significant room for LLM improvement.</abstract>
<identifier type="citekey">rahimi-etal-2026-reframe</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.2520/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>57310</start>
<end>57325</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T REFRAME: A Test Suite for the Knowledge–Deployment Gap in LLMs
%A Rahimi, Elahe
%A Roewer-Després, François
%A Sajjad, Hassan
%A Alam, Md Mahbub
%A Badshah, Sher
%A Dolatabadi, Elham
%A Rudzicz, Frank
%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 rahimi-etal-2026-reframe
%X Large language models (LLMs) are commonly evaluated on multiple-choice tasks that ask them to select an answer or to abstain on unanswerable questions. Neither setting captures frame-breaking: where the question is answerable, but the correct answer lies outside the framing of a candidate set of invalid answers. A useful model must reframe the task, rather than choose an invalid frame-compliant answer. We introduce REFRAME, a test suite of 150 scenarios in which a commonsense constraint invalidates their framing. Each REFRAME scenario is additionally paired with a direct knowledge probe of the same frame-breaking constraint, which we leverage to separate what a model knows from what it deploys. Across nine frontier LLMs, we observe a substantial knowledge-deployment gap with zero-shot prompting. A targeted reasoning prompt only partially closes this gap. Human annotators, by contrast, readily detect the constraint and reframe the problem. Together, these results characterize frame-breaking as a distinct reasoning challenge and demonstrate significant room for LLM improvement.
%U https://aclanthology.org/2026.emnlp-main.2520/
%P 57310-57325
Markdown (Informal)
[REFRAME: A Test Suite for the Knowledge–Deployment Gap in LLMs](https://aclanthology.org/2026.emnlp-main.2520/) (Rahimi et al., EMNLP 2026)
ACL
- Elahe Rahimi, François Roewer-Després, Hassan Sajjad, Md Mahbub Alam, Sher Badshah, Elham Dolatabadi, and Frank Rudzicz. 2026. REFRAME: A Test Suite for the Knowledge–Deployment Gap in LLMs. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 57310–57325, Budapest, Hungary. Association for Computational Linguistics.