Elías Abad-Rocamora, Christian Schlarmann, Naman Deep Singh, Yongtao Wu, Matthias Hein, Volkan Cevher: Robustness in Both Domains: CLIP Needs a Robust Text Encoder. NeurIPS 2025