中文
相关论文

相关论文: Cross-Modal Safety Mechanism Transfer in Large Vis…

200 篇论文

Benefiting from the powerful capabilities of Large Language Models (LLMs), pre-trained visual encoder models connected to an LLMs can realize Vision Language Models (VLMs). However, existing research shows that the visual modality of VLMs…

计算机视觉与模式识别 · 计算机科学 2024-05-24 Zhendong Liu , Yuanbi Nie , Yingshui Tan , Xiangyu Yue , Qiushi Cui , Chongjun Wang , Xiaoyong Zhu , Bo Zheng

Large visual-language models (LVLMs) integrate aligned large language models (LLMs) with visual modules to process multimodal inputs. However, the safety mechanisms developed for text-based LLMs do not naturally extend to visual modalities,…

计算机视觉与模式识别 · 计算机科学 2025-07-29 Shen Li , Liuyi Yao , Wujia Niu , Lan Zhang , Yaliang Li

Benefiting from the powerful capabilities of Large Language Models (LLMs), pre-trained visual encoder models connected to LLMs form Vision Language Models (VLMs). However, recent research shows that the visual modality in VLMs is highly…

计算机视觉与模式识别 · 计算机科学 2025-01-14 Zhendong Liu , Yuanbi Nie , Yingshui Tan , Jiaheng Liu , Xiangyu Yue , Qiushi Cui , Chongjun Wang , Xiaoyong Zhu , Bo Zheng

The emergence of vision language models (VLMs) comes with increased safety concerns, as the incorporation of multiple modalities heightens vulnerability to attacks. Although VLMs can be built upon LLMs that have textual safety alignment, it…

密码学与安全 · 计算机科学 2025-02-18 Qin Liu , Fei Wang , Chaowei Xiao , Muhao Chen

The safety alignment ability of Vision-Language Models (VLMs) is prone to be degraded by the integration of the vision module compared to its LLM backbone. We investigate this phenomenon, dubbed as ''safety alignment degradation'' in this…

Vision-Language adaptation (VL adaptation) transforms Large Language Models (LLMs) into Large Vision-Language Models (LVLMs) for multimodal tasks, but this process often compromises the inherent safety capabilities embedded in the original…

计算与语言 · 计算机科学 2024-11-18 Seongyun Lee , Geewook Kim , Jiyeon Kim , Hyunji Lee , Hoyeon Chang , Sue Hyun Park , Minjoon Seo

Multimodal large language models (MLLMs) face safety misalignment, where visual inputs enable harmful outputs. To address this, existing methods require explicit safety labels or contrastive data; yet, threat-related concepts are concrete…

计算机视觉与模式识别 · 计算机科学 2026-04-16 Qishun Yang , Shu Yang , Lijie Hu , Di Wang

Recent studies reveal that integrating new modalities into Large Language Models (LLMs), such as Vision-Language Models (VLMs), creates a new attack surface that bypasses existing safety training techniques like Supervised Fine-tuning (SFT)…

Large vision-language models (LVLMs) have achieved impressive results in various vision-language tasks. However, despite showing promising performance, LVLMs suffer from hallucinations caused by language bias, leading to diminished focus on…

计算机视觉与模式识别 · 计算机科学 2026-05-29 Haozhe Zhao , Shuzheng Si , Liang Chen , Yichi Zhang , Maosong Sun , Mingjia Zhang , Baobao Chang

Understanding the mechanisms behind Large Language Models (LLMs) is crucial for designing improved models and strategies. While recent studies have yielded valuable insights into the mechanisms of textual LLMs, the mechanisms of Multi-modal…

计算与语言 · 计算机科学 2025-01-14 Zeping Yu , Sophia Ananiadou

The deployment of multimodal large language models (MLLMs) has demonstrated remarkable success in engaging in conversations involving visual inputs, thanks to the superior power of large language models (LLMs). Those MLLMs are typically…

计算与语言 · 计算机科学 2024-10-10 Jiahui Gao , Renjie Pi , Tianyang Han , Han Wu , Lanqing Hong , Lingpeng Kong , Xin Jiang , Zhenguo Li

Large-scale Vision Language Models (LVLMs) exhibit advanced capabilities in tasks that require visual information, including object detection. These capabilities have promising applications in various industrial domains, such as autonomous…

计算机视觉与模式识别 · 计算机科学 2025-12-01 Haruki Sakajo , Hiroshi Takato , Hiroshi Tsutsui , Komei Soda , Hidetaka Kamigaito , Taro Watanabe

Ensuring Vision-Language Models (VLMs) generate safe outputs is crucial for their reliable deployment. However, LVLMs suffer from drastic safety degradation compared to their LLM backbone. Even blank or irrelevant images can trigger LVLMs…

人工智能 · 计算机科学 2025-06-02 Wenhan Yang , Spencer Stice , Ali Payani , Baharan Mirzasoleiman

Currently, inspired by the success of vision-language models (VLMs), an increasing number of researchers are focusing on improving VLMs and have achieved promising results. However, most existing methods concentrate on optimizing the…

计算机视觉与模式识别 · 计算机科学 2024-09-23 Dawei Yan , Pengcheng Li , Yang Li , Hao Chen , Qingguo Chen , Weihua Luo , Wei Dong , Qingsen Yan , Haokui Zhang , Chunhua Shen

Large vision-language models (LVLMs) are vulnerable to harmful input compared to their language-only backbones. We investigated this vulnerability by exploring LVLMs internal dynamics, framing their inherent safety understanding in terms of…

人工智能 · 计算机科学 2025-07-30 Wanying Wang , Zeyu Ma , Han Zheng , Xin Tan , Mingang Chen

Laboratories are prone to severe injuries from minor unsafe actions, yet continuous safety monitoring -- beyond mandatory pre-lab safety training -- is limited by human availability. Vision language models (VLMs) offer promise for…

计算机视觉与模式识别 · 计算机科学 2026-02-03 Trishna Chakraborty , Udita Ghosh , Aldair Ernesto Gongora , Ruben Glatt , Yue Dong , Jiachen Li , Amit K. Roy-Chowdhury , Chengyu Song

Large language models (LLMs) excel in diverse applications but face dual challenges: generating harmful content under jailbreak attacks and over-refusal of benign queries due to rigid safety mechanisms. These issues are further complicated…

人工智能 · 计算机科学 2025-11-04 Yifan Xia , Guorui Chen , Wenqian Yu , Zhijiang Li , Philip Torr , Jindong Gu

Hallucination poses a challenge to the deployment of large vision-language models (LVLMs) in applications. Unlike in large language models (LLMs), hallucination in LVLMs often arises from misalignments between visual inputs and textual…

计算机视觉与模式识别 · 计算机科学 2024-10-23 Sheng Liu , Haotian Ye , Lei Xing , James Zou

Large Vision-Language Models (LVLMs) have achieved impressive performance across multimodal understanding and reasoning tasks, yet their internal safety mechanisms remain opaque and poorly controlled. In this work, we present a…

计算机视觉与模式识别 · 计算机科学 2026-03-31 Jinhu Fu , Yihang Lou , Qingyi Si , Shudong Zhang , Yan Bai , Sen Su

Vision-language pre-training (VLP) models have shown vulnerability to adversarial examples in multimodal tasks. Furthermore, malicious adversaries can be deliberately transferred to attack other black-box models. However, existing work has…

计算机视觉与模式识别 · 计算机科学 2023-07-27 Dong Lu , Zhiqiang Wang , Teng Wang , Weili Guan , Hongchang Gao , Feng Zheng
‹ 上一页 1 2 3 10 下一页 ›