Zongqi Wang

P-GenRM: Personalized Generative Reward Model with Test-time User-based Scaling

S2J: Bridging the Gap Between Solving and Judging Ability in Generative Reward Models

Probing the robustness of large language models safety to latent perturbations

ESpeW: Robust Copyright Protection for LLM-based EaaS via Embedding-Specific Watermark