CausalVQA: A Physically Grounded Causal Reasoning Benchmark for Video Models

Used by V-JEPA 2 for benchmarking.

Citation

(Foss et al., 2025)

Foss, A., Evans, C., Mitts, S., Sinha, K., Rizvi, A., & Kao, J. T. (2025). CausalVQA: A Physically Grounded Causal Reasoning Benchmark for Video Models. https://arxiv.org/pdf/2506.09943