arXiv:2512.06674v1 Announce Type: new 
Abstract: Image-to-Video (I2V) generation synthesizes dynamic visual content from image and text inputs, providing significant creative control. However, the security of such multimodal systems, particularly their vulnerability to jailbreak attacks, remains critically underexplored. To bridge this gap, we propose RunawayEvil, the first multimodal jailbreak framework for I2V models with dynamic evolutionary capability. Built on a "Strategy-Tactic-Action" paradigm, our framework exhibits self-amplifying attack through three core components: (1) Strategy-Aware Command Unit that enables the attack to self-evolve its strategies through reinforcement learning-driven strategy customization and LLM-based strategy exploration; (2) Multimodal Tactical Planning Unit that generates coordinated text jailbreak instructions and image tampering guidelines based on the selected strategies; (3) Tactical Action Unit that executes and evaluates the multimodal coordinated attacks. This self-evolving architecture allows the framework to continuously adapt and intensify its attack strategies without human intervention. Extensive experiments demonstrate RunawayEvil achieves state-of-the-art attack success rates on commercial I2V models, such as Open-Sora 2.0 and CogVideoX. Specifically, RunawayEvil outperforms existing methods by 58.5 to 79 percent on COCO2017. This work provides a critical tool for vulnerability analysis of I2V models, thereby laying a foundation for more robust video generation systems.

يمثل تقديم RunawayEvil تقدمًا كبيرًا في مجال توليد الصور إلى فيديو (I2V)، حيث يركز على ضعف الأنظمة متعددة الوسائط أمام هجمات كسر الحماية. يستخدم هذا الإطار نموذجًا فريدًا من نوعه يسمى 'استراتيجية-تكتيك-عمل'، مما يمكّن استراتيجيات الهجوم من التطور الذاتي من خلال التعلم المعزز واستكشاف نماذج اللغة الكبيرة (LLM).

La introducción de RunawayEvil marca un avance significativo en el campo de la generación de Imágenes a Video (I2V), centrándose en las vulnerabilidades de los sistemas multimodales ante ataques de jailbreak. Este marco emplea un paradigma único de 'Estrategia-Táctica-Acción', que permite estrategias de ataque auto-evolutivas a través del aprendizaje por refuerzo y la exploración de modelos de lenguaje de gran tamaño (LLM).

L'introduction de RunawayEvil représente une avancée significative dans le domaine de la génération d'images en vidéo (I2V), en se concentrant sur les vulnérabilités des systèmes multimodaux face aux attaques de jailbreak. Ce cadre utilise un paradigme unique 'Stratégie-Tactique-Action', permettant des stratégies d'attaque auto-évolutives grâce à l'apprentissage par renforcement et à l'exploration de modèles de langage de grande taille (LLM).

The introduction of RunawayEvil marks a significant advancement in the field of Image-to-Video (I2V) generation, focusing on the vulnerabilities of multimodal systems to jailbreak attacks. This framework employs a unique 'Strategy-Tactic-Action' paradigm, enabling self-evolving attack strategies through reinforcement learning and large language model (LLM) exploration.

RunawayEvil: Jailbreaking the Image-to-Video Generative Models

arXiv:2512.06663v1 Announce Type: new 
Abstract: Large Vision-Language Models (LVLMs) have demonstrated remarkable success in a broad range of vision-language tasks, such as general visual question answering and optical character recognition (OCR). However, their performance on perception-centric tasks -- such as object detection, semantic segmentation, and depth estimation -- remains significantly inferior to that of task-specific expert models. For example, Qwen2.5-VL-7B-Instruct achieves only 19% mAP on COCO2017 val, particularly struggling with dense scenes and small object recall. In this work, we introduce Chain-of-Thought for Detection (CoT4Det), a simple but efficient strategy that reformulates perception tasks into three interpretable steps: classification, counting, and grounding -- each more naturally aligned with the reasoning capabilities of LVLMs. Extensive experiments demonstrate that our method significantly improves perception performance without compromising general vision language capabilities. With a standard Qwen2.5-VL-7B-Instruct, CoT4Det boosts mAP from 19.0% to 33.0% on COCO2017 val and achieves competitive results across a variety of perception benchmarks, outperforming baselines by +2% on RefCOCO series and 19% on Flickr30k entities.

تهدف مقدمة CoT4Det، وهو إطار تفكير متسلسل، إلى تحسين أداء نماذج اللغة والرؤية الكبيرة (LVLM) في المهام الموجهة نحو الإدراك مثل اكتشاف الكائنات والتجزئة الدلالية، والتي كانت سابقًا متخلفة عن النماذج المتخصصة في المهام. يعيد هذا الإطار صياغة هذه المهام في ثلاث خطوات قابلة للتفسير: التصنيف، العد، والتثبيت.

La introducción de CoT4Det, un marco de pensamiento encadenado, tiene como objetivo mejorar el rendimiento de los Modelos de Lenguaje-Visión de Gran Escala (LVLM) en tareas orientadas a la percepción, como la detección de objetos y la segmentación semántica, que anteriormente habían quedado rezagadas en comparación con los modelos específicos para tareas. Este marco reformula estas tareas en tres pasos interpretables: clasificación, conteo y anclaje.

L'introduction de CoT4Det, un cadre de réflexion en chaîne, vise à améliorer les performances des grands modèles de vision-langage (LVLM) sur des tâches orientées vers la perception telles que la détection d'objets et la segmentation sémantique, qui ont précédemment été en retard par rapport aux modèles spécifiques à des tâches. Ce cadre reformule ces tâches en trois étapes interprétables : classification, comptage et ancrage.

The introduction of CoT4Det, a Chain-of-Thought framework, aims to enhance the performance of Large Vision-Language Models (LVLMs) on perception-oriented tasks such as object detection and semantic segmentation, which have previously lagged behind task-specific models. This framework reformulates these tasks into three interpretable steps: classification, counting, and grounding.

RunawayEvil: Jailbreaking the Image-to-Video Generative Models

Was this article worth reading? Share it

SwapAnything.io

Republiclabs.ai

AiReelGenerator.com