Search Results for author:
Found 3 papers, 1 papers with code
Improving the Robustness of Transformer-based Large Language Models with Dynamic Attention
Extensive experiments demonstrate that dynamic attention significantly mitigates the impact of adversarial attacks, improving up to 33\% better performance than previous methods against widely-used adversarial attacks.
TextDefense: Adversarial Text Detection based on Word Importance Entropy
TextDefense differs from previous approaches, where it utilizes the target model for detection and thus is attack type agnostic.
Backdoor Pre-trained Models Can Transfer to All
However, a pre-trained model with backdoor can be a severe threat to the applications.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
