amazon-agi/Amazon-Nova-1.0-Premier-evals

Amazon Nova Premier Evaluation数据集包含多个子数据集，用于评估Nova Premier模型在各种NLP任务上的性能。这些任务包括MMLU、GPQA-Diamond、AIME 2025、Math-500、BigCodeBench、IFEval、MMMU、CharXiv、EgoSchema、TallyQA、Screenspot和BFCL等。每个子数据集都有不同的应用场景和评估方式。The Amazon Nova Premier Evaluation dataset consists of multiple sub-datasets designed to assess the performance of the Nova Premier model on various NLP tasks. These tasks include MMLU, GPQA-Diamond, AIME 2025, Math-500, BigCodeBench, IFEval, MMMU, CharXiv, EgoSchema, TallyQA, Screenspot, and BFCL. Each sub-dataset has its specific application scenario and evaluation method.