1 paper · 1 filter
Shuyue Stella Li, Jacqueline He, Jiahui Chen +5
Video reward models are trained and evaluated on human preferences between videos that often differ simultaneously on properties such as prompt alignment, visual fidelity, motion,…