davidanugraha/rStarCoder-Human-Test-4k

该数据集包含4000个训练样本，每个样本具有多个特征，包括id、data_source、content、ability、reward_model、extra_info和prompt。其中，reward_model包含ground_truth和style子特征，ground_truth进一步包含input和output结构。extra_info包含func_name、id和starter_code。prompt是一个列表，包含content和role。数据集的总大小为2719524479字节。

This dataset contains 4000 training samples, each with multiple features including id, data_source, content, ability, reward_model, extra_info, and prompt. The reward_model feature includes sub-features ground_truth and style, with ground_truth further structured into input and output. extra_info includes func_name, id, and starter_code. prompt is a list containing content and role. The total size of the dataset is 2719524479 bytes.