Minimal, Local, Causal Explanations for Jailbreak Success in Large Language Models

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Kumar, Shubham, Ahuja, Narendra
Formato:	Preprint
Publicado:	2026
Materias:	Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Measuring the (Un)Faithfulness of Concept-Based Explanations
por: Kumar, Shubham, et al.
Publicado: (2025)

LUQ: Layerwise Ultra-Low Bit Quantization for Multimodal Large Language Models
por: Bhatnagar, Shubhang, et al.
Publicado: (2025)

Locally-Minimal Probabilistic Explanations
por: Izza, Yacine, et al.
Publicado: (2023)

GPT-4 Jailbreaks Itself with Near-Perfect Success Using Self-Explanation
por: Ramesh, Govind, et al.
Publicado: (2024)

Causality-Aware Local Interpretable Model-Agnostic Explanations
por: Cinquini, Martina, et al.
Publicado: (2022)

Understanding Jailbreak Success: A Study of Latent Space Dynamics in Large Language Models
por: Ball, Sarah, et al.
Publicado: (2024)

Piecewise-Linear Manifolds for Deep Metric Learning
por: Bhatnagar, Shubhang, et al.
Publicado: (2024)

Boosting Jailbreak Transferability for Large Language Models
por: Liu, Hanqing, et al.
Publicado: (2024)

EasyJailbreak: A Unified Framework for Jailbreaking Large Language Models
por: Zhou, Weikang, et al.
Publicado: (2024)

Potential Field Based Deep Metric Learning
por: Bhatnagar, Shubhang, et al.
Publicado: (2024)

Challenges and Applications of Large Language Models: A Comparison of GPT and DeepSeek family of models
por: Sharma, Shubham, et al.
Publicado: (2025)

Jailbreaking Large Language Models with Symbolic Mathematics
por: Bethany, Emet, et al.
Publicado: (2024)

Single-Configuration Attack Success Rate Is Not Enough: Jailbreak Evaluations Should Report Distributional Attack Success
por: Maple, Carsten, et al.
Publicado: (2026)

Emoji-Based Jailbreaking of Large Language Models
por: Gopinadh, M P V S, et al.
Publicado: (2026)

Jailbreaking Large Vision Language Models in Intelligent Transportation Systems
por: Das, Badhan Chandra, et al.
Publicado: (2025)

Knowledge-to-Jailbreak: Investigating Knowledge-driven Jailbreaking Attacks for Large Language Models
por: Tu, Shangqing, et al.
Publicado: (2024)

BiasJailbreak:Analyzing Ethical Biases and Jailbreak Vulnerabilities in Large Language Models
por: Lee, Isack, et al.
Publicado: (2024)

Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models
por: Song, Zirui, et al.
Publicado: (2025)

Improving Tool Retrieval by Leveraging Large Language Models for Query Generation
por: Kachuee, Mohammad, et al.
Publicado: (2024)

Jailbreaking and Mitigation of Vulnerabilities in Large Language Models
por: Peng, Benji, et al.
Publicado: (2024)

Imperceptible Jailbreaking against Large Language Models
por: Gao, Kuofeng, et al.
Publicado: (2025)

Large Causal Models from Large Language Models
por: Mahadevan, Sridhar
Publicado: (2025)

A Cross-Language Investigation into Jailbreak Attacks in Large Language Models
por: Li, Jie, et al.
Publicado: (2024)

Rethinking Prompting Strategies for Multi-Label Recognition with Partial Annotations
por: Rawlekar, Samyak, et al.
Publicado: (2024)

GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts
por: Yu, Jiahao, et al.
Publicado: (2023)

Evaluating Multi-Agent Defences Against Jailbreaking Attacks on Large Language Models
por: Wit, Maria Carolina Cornelia, et al.
Publicado: (2025)

ShallowJail: Steering Jailbreaks against Large Language Models
por: Liu, Shang, et al.
Publicado: (2026)

The Cost of Thinking: Increased Jailbreak Risk in Large Language Models
por: Yang, Fan
Publicado: (2025)

Behind the Mask: Benchmarking Camouflaged Jailbreaks in Large Language Models
por: Zheng, Youjia, et al.
Publicado: (2025)

Jailbreaking Black Box Large Language Models in Twenty Queries
por: Chao, Patrick, et al.
Publicado: (2023)

SoK: Evaluating Jailbreak Guardrails for Large Language Models
por: Wang, Xunguang, et al.
Publicado: (2025)

Causal Explanations for Image Classifiers
por: Chockler, Hana, et al.
Publicado: (2024)

A Framework for Causal Concept-based Model Explanations
por: Bjøru, Anna Rodum, et al.
Publicado: (2025)

JailbreakEval: An Integrated Toolkit for Evaluating Jailbreak Attempts Against Large Language Models
por: Ran, Delong, et al.
Publicado: (2024)

Large Language Models as Nondeterministic Causal Models
por: Beckers, Sander
Publicado: (2025)

Local Success Does Not Compose: Benchmarking Large Language Models for Compositional Formal Verification
por: Xu, Xu, et al.
Publicado: (2025)

Jailbreaking Large Language Models Through Content Concretization
por: Wahréus, Johan, et al.
Publicado: (2025)

Distract Large Language Models for Automatic Jailbreak Attack
por: Xiao, Zeguan, et al.
Publicado: (2024)

Defending Large Language Models Against Jailbreak Attacks via Layer-specific Editing
por: Zhao, Wei, et al.
Publicado: (2024)

SoK: Robustness in Large Language Models against Jailbreak Attacks
por: Xu, Feiyue, et al.
Publicado: (2026)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
VUFIND_SESSION	"d6d769440bbc579041b3e184234dee59"
language	"es"
ui	"standard"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790827527.1913 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "d6d769440bbc579041b3e184234dee59" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"VUFIND_SESSION=d6d769440bbc579041b3e184234dee59; language=es; ui=standard"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.216.147"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"54764"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2605_00123/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2605_00123/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790827527.1868
REQUEST_TIME	1790827527

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler