From LLMs to MLLMs: Exploring the Landscape of Multimodal Jailbreaking
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929393732943872 |
|---|---|
| author | Wang, Siyuan Long, Zhuohan Fan, Zhihao Wei, Zhongyu |
| author_facet | Wang, Siyuan Long, Zhuohan Fan, Zhihao Wei, Zhongyu |
| contents | The rapid development of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has exposed vulnerabilities to various adversarial attacks. This paper provides a comprehensive overview of jailbreaking research targeting both LLMs and MLLMs, highlighting recent advancements in evaluation benchmarks, attack techniques and defense strategies. Compared to the more advanced state of unimodal jailbreaking, multimodal domain remains underexplored. We summarize the limitations and potential research directions of multimodal jailbreaking, aiming to inspire future research and further enhance the robustness and security of MLLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14859 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | From LLMs to MLLMs: Exploring the Landscape of Multimodal Jailbreaking Wang, Siyuan Long, Zhuohan Fan, Zhihao Wei, Zhongyu Computation and Language Artificial Intelligence The rapid development of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has exposed vulnerabilities to various adversarial attacks. This paper provides a comprehensive overview of jailbreaking research targeting both LLMs and MLLMs, highlighting recent advancements in evaluation benchmarks, attack techniques and defense strategies. Compared to the more advanced state of unimodal jailbreaking, multimodal domain remains underexplored. We summarize the limitations and potential research directions of multimodal jailbreaking, aiming to inspire future research and further enhance the robustness and security of MLLMs. |
| title | From LLMs to MLLMs: Exploring the Landscape of Multimodal Jailbreaking |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2406.14859 |