Below we show a small subset of OmniBench subtasks, which allows you to explore the data in detail.
—
def evaluate_task(): # TODO: Implement evaluation logic return "Not implemented"