EIBench

该数据集是一个大规模的基准测试，包含了1615个基础情感解释（EI）样本和50个复杂情感解释样本，这些样本展现了多方面的情感，并专注于驱动情感反应的各种因果因素。每个实例都需要基于理由的解释，而不是简单的分类，同时支持多种评估模式，以评估语言模型的能力。该任务被称为情感解释（EI）。This dataset is a large-scale benchmark containing 1,615 basic emotion interpretation (EI) samples and 50 complex emotion interpretation samples. These samples cover a wide range of emotions and focus on diverse causal factors that drive emotional responses. Each instance requires justification-based explanations rather than simple classification, and supports multiple evaluation paradigms to assess the capabilities of language models. This task is termed Emotion Interpretation (EI).