Login / Signup

TransTroj: Transferable Backdoor Attacks to Pre-trained Models via Embedding Indistinguishability.

Hao WangTao XiangShangwei GuoJialing HeHangcheng LiuTianwei Zhang
Published in: CoRR (2024)
Keyphrases
  • pre trained
  • high dimensional
  • text classification
  • statistical model