CycliST

CycliST: A Video Language Model Benchmark for Reasoning on Cyclical State Transitions CycliST is a synthetic, diagnostic benchmark for evaluating Video Language Models (VLMs) on their ability to reason over cyclical state transitions; periodic patterns in object motion and visual attributes. Published in the Journal of Data-centric Machine Learning Research (DMLR), 2026. Dataset Summary Cyclical patterns are everywhere in the physical world, from traffic… See the full description on the dataset page:

Source: Hugging Face Hub (AIML-TUDA/CycliST). Metadata imported from the dataset’s Hub tags.