Mitigating Gradient Inversion Risks in Language Models via Token Obfuscation
arXiv:2602.15897v1 Announce Type: new Abstract: Training and fine-tuning large-scale language models largely benefit from collaborative learning, but the approach has been proven vulnerable to gradient …
Xinguo Feng, Zhongkui Ma, Zihan Wang, Alsharif Abuadbba, Guangdong Bai
6 views
