【论文】When Does Latent Prediction Error Support Physical Surprise Claims in Video World Models?

arXiv 2611.01815 · 20 人推荐 · 附代码 3-ml/jepa-surprise。Prediction errors from video world models are often treated as physical surprise, but they can also reflect appearance, static content, motion, or ordinary prediction difficulty. We present an evaluation framework that separates what frozen representations contain, how they organize appearance and physical differences, and what changes a model's own prediction error. On IntPhys2, a benchmark of physically possible and impossible events, we compare representations from latent prediction, masked p