[ECCV 2026] VKnowU: Evaluating Visual Knowledge Understanding in Multimodal LLMs
reinforcement-learning video-understanding commonsense-reasoning visual-reasoning intuitive-physics video-question-answering video-benchmark human-centric mllm multimodal-llm world-knowledge video-llm visual-knowledge vision-language-m
-
Updated
Feb 3, 2026 - Python