@inproceedings{oskooei-aktas-2026-breakfun,
title = "{B}reak{F}un: Jailbreaking {LLM}s via Object Instantiation under Simulated Code Execution",
author = "Oskooei, Amirkia Rafiei and
Aktas, Mehmet",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.52/",
pages = "1019--1038",
ISBN = "979-8-89176-444-6",
abstract = "Large Language Models (LLMs) are widely used because they process structures, syntax and code well, but this same ability also makes them paradoxically vulnerable. We introduce BreakFun, a jailbreak method that frames a harmful request as code-execution simulation. The prompt gives the model a benign Python class definition, the ``Trojan Schema'', and asks what that code would print if it ran. To answer, the model must create an object from the class and invent a value for each field, and the adversarial field names steer those values toward the attacker{'}s goal. The harmful content therefore appears as a side-effect of the simulated instantiation, not as a direct answer. A three-part prompt carries the attack: an innocent frame, the Trojan Schema, and a Chain-of-Thought distraction. On JailbreakBench, BreakFun reaches an average attack success rate of 89{\%} across 13 open-weight and commercial models (open-weight {\textasciitilde}98{\%}, API systems {\textasciitilde}78{\%}), and 100{\%} on several. An ablation identifies the Trojan Schema as the strongest contributing component. As a counter-ablation probe we introduce Adversarial Prompt Deconstruction (APD), in which a second LLM writes out all readable text in the prompt before it judges safety. APD improves detection across three model families, and a probe-specific ablation shows that the transcription step gives most of the gain. Our findings show that LLM safety must account for content that a simulation task produces as a side-effect, not only for content that a user requests directly."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="oskooei-aktas-2026-breakfun">
<titleInfo>
<title>BreakFun: Jailbreaking LLMs via Object Instantiation under Simulated Code Execution</title>
</titleInfo>
<name type="personal">
<namePart type="given">Amirkia</namePart>
<namePart type="given">Rafiei</namePart>
<namePart type="family">Oskooei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mehmet</namePart>
<namePart type="family">Aktas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>Large Language Models (LLMs) are widely used because they process structures, syntax and code well, but this same ability also makes them paradoxically vulnerable. We introduce BreakFun, a jailbreak method that frames a harmful request as code-execution simulation. The prompt gives the model a benign Python class definition, the “Trojan Schema”, and asks what that code would print if it ran. To answer, the model must create an object from the class and invent a value for each field, and the adversarial field names steer those values toward the attacker’s goal. The harmful content therefore appears as a side-effect of the simulated instantiation, not as a direct answer. A three-part prompt carries the attack: an innocent frame, the Trojan Schema, and a Chain-of-Thought distraction. On JailbreakBench, BreakFun reaches an average attack success rate of 89% across 13 open-weight and commercial models (open-weight ~98%, API systems ~78%), and 100% on several. An ablation identifies the Trojan Schema as the strongest contributing component. As a counter-ablation probe we introduce Adversarial Prompt Deconstruction (APD), in which a second LLM writes out all readable text in the prompt before it judges safety. APD improves detection across three model families, and a probe-specific ablation shows that the transcription step gives most of the gain. Our findings show that LLM safety must account for content that a simulation task produces as a side-effect, not only for content that a user requests directly.</abstract>
<identifier type="citekey">oskooei-aktas-2026-breakfun</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.52/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>1019</start>
<end>1038</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T BreakFun: Jailbreaking LLMs via Object Instantiation under Simulated Code Execution
%A Oskooei, Amirkia Rafiei
%A Aktas, Mehmet
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F oskooei-aktas-2026-breakfun
%X Large Language Models (LLMs) are widely used because they process structures, syntax and code well, but this same ability also makes them paradoxically vulnerable. We introduce BreakFun, a jailbreak method that frames a harmful request as code-execution simulation. The prompt gives the model a benign Python class definition, the “Trojan Schema”, and asks what that code would print if it ran. To answer, the model must create an object from the class and invent a value for each field, and the adversarial field names steer those values toward the attacker’s goal. The harmful content therefore appears as a side-effect of the simulated instantiation, not as a direct answer. A three-part prompt carries the attack: an innocent frame, the Trojan Schema, and a Chain-of-Thought distraction. On JailbreakBench, BreakFun reaches an average attack success rate of 89% across 13 open-weight and commercial models (open-weight ~98%, API systems ~78%), and 100% on several. An ablation identifies the Trojan Schema as the strongest contributing component. As a counter-ablation probe we introduce Adversarial Prompt Deconstruction (APD), in which a second LLM writes out all readable text in the prompt before it judges safety. APD improves detection across three model families, and a probe-specific ablation shows that the transcription step gives most of the gain. Our findings show that LLM safety must account for content that a simulation task produces as a side-effect, not only for content that a user requests directly.
%U https://aclanthology.org/2026.findings-aacl.52/
%P 1019-1038
Markdown (Informal)
[BreakFun: Jailbreaking LLMs via Object Instantiation under Simulated Code Execution](https://aclanthology.org/2026.findings-aacl.52/) (Oskooei & Aktas, Findings 2026)
ACL