一些对抗领域的论文复现
(1)[FAST IS BETTER THAN FREE: REVISITING ADVERSARIAL TRAINING] https://arxiv.org/pdf/2001.03994v1.pdf
(2)[Universal Adversarial Triggers for Attacking and Analyzing NLP] https://arxiv.org/pdf/1908.07125v3.pdf
(3)[GPT-4 IS TOO SMART TO BE SAFE: STEALTHY CHAT WITH LLMS VIA CIPHER]https://arxiv.org/pdf/2308.06463.pdf