kothasuhas/llama-3b-gold-15M-student-generations_SNIS_2048_tune422v1_N15.00M_T16.0

这是一个包含文本数据的 dataset，数据集字段包括文本内容（text），对数权重（log_weight），抽样概率缩放（sampling_p_scaled）和抽样温度缩放（sampling_p_temperature_scaled）。数据集分为训练集和验证集，训练集有1500万个示例，验证集有1000个示例。

This is a dataset containing text data, with fields including text content (text), logarithmic weight (log_weight), scaled sampling probability (sampling_p_scaled), and scaled sampling temperature (sampling_p_temperature_scaled). The dataset is split into training and validation sets, with the training set containing 15 million examples and the validation set containing 1,000 examples.