five

andrewsiah/MetaLearningRewardDatabase1K

收藏
Hugging Face2024-06-06 更新2024-06-12 收录
下载链接:
https://hf-mirror.com/datasets/andrewsiah/MetaLearningRewardDatabase1K
下载链接
链接失效反馈
官方服务:
资源简介:
--- dataset_info: features: - name: person_weight sequence: float64 - name: prompt_1 dtype: string - name: response_1 dtype: string - name: reward_1 dtype: float64 - name: prompt_2 dtype: string - name: response_2 dtype: string - name: reward_2 dtype: float64 - name: prompt_3 dtype: string - name: response_3 dtype: string - name: reward_3 dtype: float64 - name: prompt_4 dtype: string - name: response_4 dtype: string - name: reward_4 dtype: float64 - name: prompt_5 dtype: string - name: response_5 dtype: string - name: reward_5 dtype: float64 - name: prompt_6 dtype: string - name: response_6 dtype: string - name: reward_6 dtype: float64 - name: prompt_7 dtype: string - name: response_7 dtype: string - name: reward_7 dtype: float64 - name: prompt_8 dtype: string - name: response_8 dtype: string - name: reward_8 dtype: float64 - name: prompt_9 dtype: string - name: response_9 dtype: string - name: reward_9 dtype: float64 - name: prompt_10 dtype: string - name: response_10 dtype: string - name: reward_10 dtype: float64 - name: prompt_11 dtype: string - name: response_11 dtype: string - name: reward_11 dtype: float64 - name: prompt_12 dtype: string - name: response_12 dtype: string - name: reward_12 dtype: float64 - name: prompt_13 dtype: string - name: response_13 dtype: string - name: reward_13 dtype: float64 - name: prompt_14 dtype: string - name: response_14 dtype: string - name: reward_14 dtype: float64 - name: prompt_15 dtype: string - name: response_15 dtype: string - name: reward_15 dtype: float64 - name: prompt_16 dtype: string - name: response_16 dtype: string - name: reward_16 dtype: float64 - name: prompt_17 dtype: string - name: response_17 dtype: string - name: reward_17 dtype: float64 - name: prompt_18 dtype: string - name: response_18 dtype: string - name: reward_18 dtype: float64 - name: prompt_19 dtype: string - name: response_19 dtype: string - name: reward_19 dtype: float64 - name: prompt_20 dtype: string - name: response_20 dtype: string - name: reward_20 dtype: float64 - name: prompt_21 dtype: string - name: response_21 dtype: string - name: reward_21 dtype: float64 - name: prompt_22 dtype: string - name: response_22 dtype: string - name: reward_22 dtype: float64 - name: prompt_23 dtype: string - name: response_23 dtype: string - name: reward_23 dtype: float64 - name: prompt_24 dtype: string - name: response_24 dtype: string - name: reward_24 dtype: float64 - name: prompt_25 dtype: string - name: response_25 dtype: string - name: reward_25 dtype: float64 - name: prompt_26 dtype: string - name: response_26 dtype: string - name: reward_26 dtype: float64 - name: prompt_27 dtype: string - name: response_27 dtype: string - name: reward_27 dtype: float64 - name: prompt_28 dtype: string - name: response_28 dtype: string - name: reward_28 dtype: float64 - name: prompt_29 dtype: string - name: response_29 dtype: string - name: reward_29 dtype: float64 - name: prompt_30 dtype: string - name: response_30 dtype: string - name: reward_30 dtype: float64 - name: prompt_31 dtype: string - name: response_31 dtype: string - name: reward_31 dtype: float64 - name: prompt_32 dtype: string - name: response_32 dtype: string - name: reward_32 dtype: float64 - name: prompt_33 dtype: string - name: response_33 dtype: string - name: reward_33 dtype: float64 - name: length dtype: float64 - name: prompt_34 dtype: string - name: response_34 dtype: string - name: reward_34 dtype: float64 - name: prompt_35 dtype: string - name: response_35 dtype: string - name: reward_35 dtype: float64 - name: prompt_36 dtype: string - name: response_36 dtype: string - name: reward_36 dtype: float64 - name: prompt_37 dtype: string - name: response_37 dtype: string - name: reward_37 dtype: float64 - name: prompt_38 dtype: string - name: response_38 dtype: string - name: reward_38 dtype: float64 - name: prompt_39 dtype: string - name: response_39 dtype: string - name: reward_39 dtype: float64 - name: prompt_40 dtype: string - name: response_40 dtype: string - name: reward_40 dtype: float64 - name: prompt_41 dtype: string - name: response_41 dtype: string - name: reward_41 dtype: float64 - name: prompt_42 dtype: string - name: response_42 dtype: string - name: reward_42 dtype: float64 - name: prompt_43 dtype: string - name: response_43 dtype: string - name: reward_43 dtype: float64 - name: prompt_44 dtype: string - name: response_44 dtype: string - name: reward_44 dtype: float64 - name: prompt_45 dtype: string - name: response_45 dtype: string - name: reward_45 dtype: float64 - name: prompt_46 dtype: string - name: response_46 dtype: string - name: reward_46 dtype: float64 - name: prompt_47 dtype: string - name: response_47 dtype: string - name: reward_47 dtype: float64 - name: prompt_48 dtype: string - name: response_48 dtype: string - name: reward_48 dtype: float64 - name: prompt_49 dtype: string - name: response_49 dtype: string - name: reward_49 dtype: float64 - name: person_id dtype: int64 splits: - name: train num_bytes: 74224100 num_examples: 1000 download_size: 45163147 dataset_size: 74224100 configs: - config_name: default data_files: - split: train path: data/train-* ---

数据集信息: 特征字段: - 名称:人员权重(person_weight),类型为序列(sequence),数据格式为64位浮点型(float64) - 名称:提示词1(prompt_1),数据类型为字符串(string) - 名称:回复1(response_1),数据类型为字符串(string) - 名称:奖励得分1(reward_1),数据格式为64位浮点型(float64) - 名称:提示词2(prompt_2),数据类型为字符串(string) - 名称:回复2(response_2),数据类型为字符串(string) - 名称:奖励得分2(reward_2),数据格式为64位浮点型(float64) 后续依次包含提示词3至提示词49、回复3至回复49、奖励得分3至奖励得分49,各组字段的数据类型均与前两组一致; - 名称:序列长度(length),数据格式为64位浮点型(float64) - 名称:人员ID(person_id),数据类型为64位整数(int64) 数据集划分: - 划分名称:训练集(train),占用字节数:74224100,样本总量:1000 下载大小:45163147字节,数据集实际占用大小:74224100字节 配置项: - 配置名称:默认配置(default) 数据文件: - 划分集:训练集(train),文件路径:data/train-*
提供机构:
andrewsiah
原始信息汇总

数据集概述

数据集特征

  • person_weight: 数据类型为float64
  • prompt_n, response_n, reward_n (n=1至49): 每个特征的数据类型为stringfloat64
  • length: 数据类型为float64
  • person_id: 数据类型为int64

数据集分割

  • train: 包含1000个示例,数据大小为74224100字节。

数据集大小

  • 下载大小: 45163147字节。
  • 数据集大小: 74224100字节。

配置

  • config_name: default
  • data_files:
    • split: train
    • path: data/train-*
二维码
社区交流群
二维码
科研交流群
商业服务