Seahorse is a dataset for multilingual, multi-faceted summarization evaluation. It consists of 96K summaries with human ratings along 6 quality dimensions: comprehensibility, repetition, grammar, attribution, main idea(s), and conciseness, covering 6 languages, 9 systems and 4 datasets.
One gem_id in the training dataset equals NaN. Its index is 14480, if counting from zero. Which gem_id does this dataset correspond to? Can you reconstruct
Hello!
One gem_id in the training dataset equals NaN. Its index is 14480, if counting from zero. Which gem_id does this dataset correspond to? Can you reconstruct