KnutJaegersberg/summeval_pairs

Dataset paired from here: https://github.com/Yale-LILY/SummEval It's smaller than I thought. Perhaps one can squeeze out a few hundred comparisons for an llm.