Textual Response-Aware Guidance for Zero-Shot Adversarial Robustness of Vision-Language Models
Text encoder responses guide robust fine-tuning of vision-language models.
Explainer · on the wire
The explainer for this story is still on the wire.
The claim, the mechanism, and the figures land here once generated.
Source · arXiv