MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation

Add code
Jan 06, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: