Xiaoyun Wang — accepted AI-conference papers

2025

FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts

Yichen Gong, Delong Ran, Jinyuan Liu, Conglei Wang, Tianshuo Cong, Anyu Wang +2

AAAI 2025technical

Large Vision-Language Models (LVLMs) signify a groundbreaking paradigm shift within the Artificial Intelligence (AI) community, extending beyond the capabilities of Large Language Models (LLMs) by assimilating additional modalities (e.g., images). Despite this advancement, the safety of LVLMs remain…

Cited by 146 Source ↗PDF ↗code ↗Scholar ↗

2020

Energy-based Out-of-distribution Detection

Weitang Liu, Xiaoyun Wang, John Owens, Yixuan Li

NeurIPS 2020poster

Determining whether inputs are out-of-distribution (OOD) is an essential building block for safely deploying machine learning models in the open world. However, previous methods relying on the softmax confidence score suffer from overconfident posterior distributions for OOD data. We propose a unifi…

Cited by 1700 Source ↗PDF ↗code ↗Scholar ↗