Author

Negar Rostamzadeh

h-index 24 View on Semantic Scholar

We have 1 of 74 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Preprint Aug 2026

TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models

A lightweight Text Encoder Alignment framework that fine-tunes only the text encoder while keeping the generative backbone fully frozen, and achieves state-of-the-art erasure robustness against black-box and white-box adversarial attacks on Stable Diffusion v1.4, while preserving generation quality on benign prompts.

Alireza Dehghanpour Farashah, Zhuan Shi, Negar Rostamzadeh et al. · 0 citations