Kamyar-zeinalipour/gemma_kg_text

该数据集包含了与文本生成相关的多个特征，如循环次数、温度、文本生成的概率、生成的原始文本、提取的输出文本、应用的模板文本、ROUGE评分、文本和三元组的ROUGE-L F度量、嵌入相似度、组合相似度、组合相似度差异、输入文本、初始文本、源文件、用户内容、助手输出以及消息内容。数据集分为训练集和测试集，共有1950个训练样本和50个测试样本。

The dataset contains multiple features related to text generation, such as cycle count, temperature, probability of text generation, raw generated text, extracted output text, applied template text, ROUGE scores, ROUGE-L F measure for text and triples, embedding similarity, combined similarity, difference in combined similarity, input text, initial text, source file, user content, assistant output, and message content. The dataset is split into a training set and a test set, with a total of 1950 training samples and 50 test samples.