shivank21/bbh_subset_llama_base

该数据集包含输入(input)、目标(target)、源配置(source_config)和模型响应(model_response)四个字符串类型的字段。它有一个训练集(train)，共有500个示例，数据集总大小为233103字节，下载大小为86444字节。数据集使用默认配置，训练数据文件的路径为data/train-*。The dataset includes four string-type fields: input, target, source_config, and model_response. It has a training set (train) with 500 examples, with a total size of 233103 bytes and a download size of 86444 bytes. The dataset uses a default configuration, with the training data file path specified as data/train-*.