TY - RPRT TI - How to Enhance Downstream Adversarial Robustness (almost) without Touching the Pre-Trained Foundation Model? AU - Meiqi Liu AU - Zhuoqun Huang AU - Yue Xing PY - 2025 UR - https://arxiv.org/abs/2504.10850 ID - 2504.10850 ER -