13 citations · 33 across the 14 of their papers we have counts for
4 papers · 1 filter
DreamDistribution: Learning Prompt Distribution for Diverse In-distribution Generation
Brian Nlong Zhao, Yuhang Xiao, Jiashu Xu +6
The popularization of Text-to-Image (T2I) diffusion models enables the generation of high-quality images from text descriptions. However, generating diverse customized images with…
Test-time Backdoor Mitigation for Black-Box Large Language Models with Defensive Demonstrations
Wenjie Mo, Jiashu Xu, Qin Liu +5
Existing studies in backdoor defense have predominantly focused on the training phase, overlooking the critical aspect of testing time defense. This gap becomes pronounced in the c…
Beyond Generation: Harnessing Text to Image Models for Object Detection and Segmentation
Yunhao Ge, Jiashu Xu, Brian Nlong Zhao +3
We propose a new paradigm to automatically generate training data with accurate labels at scale using the text-to-image synthesis frameworks (e.g., DALL-E, Stable Diffusion, etc.).…
Instructions as Backdoors: Backdoor Vulnerabilities of Instruction Tuning for Large Language Models
Jiashu Xu, Mingyu Derek Ma, Fei Wang +2
We investigate security concerns of the emergent instruction tuning paradigm, that models are trained on crowdsourced datasets with task instructions to achieve superior performanc…