#ai-detector-evasion

RoBERTa 기반 Reward 설계를 통한 AI Detector TPR 87.88% 감소

Notes on adversarial paraphrasing: a paper review