ParsiNLU, ArmanEmo, ArmanNER, ConjNLI, MATH
收藏资源简介:
本研究使用多个数据集,包括ParsiNLU、ArmanEmo、ArmanNER等,以及自创的数学和逻辑问题数据集,总计超过100万条数据。这些数据集用于评估大型语言模型在波斯语环境下的性能,涉及情感分析、阅读理解、翻译等多种任务。数据集的创建和选择考虑了任务的相关性、难度、多样性和质量,旨在全面评估模型在波斯语环境中的表现,特别是在数学和逻辑推理方面的能力。
This study utilizes multiple datasets, including ParsiNLU, ArmanEmo, ArmanNER, and others, as well as a self-developed dataset of mathematical and logical reasoning problems, totaling over 1 million data samples. These datasets are employed to evaluate the performance of large language models in the Persian language context, covering a range of tasks such as sentiment analysis, reading comprehension, machine translation, and more. The creation and selection of these datasets take into account task relevance, difficulty, diversity, and quality, with the goal of comprehensively assessing the model's performance in the Persian language scenario, particularly its mathematical and logical reasoning capabilities.




