Backdoor Attacks in Token Selection of Attention Mechanism

Yunjuan Wang; Raman Arora

2025 ICML ICML 2025

Backdoor Attacks in Token Selection of Attention Mechanism

Abstract

No abstract available. No abstract was captured for this paper when its proceedings were scraped - most often an older edition published without one online. How the corpus is built.

Authors

Yunjuan Wang, Raman Arora

Topics

Artificial Intelligence > Core AI > Foundation Models Machine Learning > Optimization & Theory > Learning Theory Deep Learning > Architectures > Transformers

Keywords

attention mechanism token selection gradient descent backdoor attack poisoned training data single-head self-attention

Download PDF