Awesome World Model Hub Papers · Datasets · Projects
← Back to datasets

Thinking in Video

2026 Datasets
Preview for Thinking in Video

Dataset Analysis

Proposes Causal-Generative Dual-Judge (CGDJ) to evaluate if video generators truly reason about real-world dynamics, revealing a perception-prediction gap.

Provenance

Collected from papers.

Derived from paper: Thinking in Video: Can Video Generators Really Reason About the Real World?

Resources