InsultedByMathematics/llama3-ultrafeedback-armo-test-evaluation-update-201

该数据集包含多个特征，主要涉及微调后的响应、提示、奖励值、选择的响应、拒绝的响应等。数据集的特征包括字符串类型的微调响应、提示、选择的响应和拒绝的响应，以及浮点数类型的奖励值和序列类型的提示标记。数据集的分割为test_prefs，包含1690个示例，总大小为105053218字节。

This dataset contains multiple features, mainly involving fine-tuned responses, prompts, reward values, chosen responses, and rejected responses. The features of the dataset include string-type fine-tuned responses, prompts, chosen responses, and rejected responses, as well as float-type reward values and sequence-type prompt tokens. The dataset is split into test_prefs, containing 1690 examples with a total size of 105053218 bytes.