ShengYun Peng

interactive visualization (2) model interpretability (2) text generation (1) image captioning (1) vision-language alignment (1) model analysis (1) saliency map (1) adversarial attack (1) generative model (1) diffusion model (1) training data attribution (1) multimodal large language model (1) inference optimization (1) model scaling (1) stable diffusion (1) adversarial example (1) video language model (1) model interpretation (1) faster r-cnn (1) llm safety (1)