Acquiring Clean Language Models from Backdoor Poisoned Datasets by Downscaling Frequency Space

Add code
Feb 27, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: