MINERVA-Cultural is a multicultural multilingual video reasoning benchmark with native-language human annotations and evidence graphs that shows state-of-the-art video LLMs perform far below human accuracy, mainly due to failures in perceiving cultural visual elements.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
MINERVA-Cultural: A Benchmark for Cultural and Multilingual Long Video Reasoning
MINERVA-Cultural is a multicultural multilingual video reasoning benchmark with native-language human annotations and evidence graphs that shows state-of-the-art video LLMs perform far below human accuracy, mainly due to failures in perceiving cultural visual elements.