MS student, Korea Advanced Institute of Science & Technology
1 paper at NeurIPS 2025
We propose a training-free safe generation method to guide text embeddings for safe text-to-image diffusion models.
