adamxyang/1.4b-policy_preference_data_gold_labelled_noisy

--- dataset_info: features: - name: answers sequence: string - name: input dtype: string - name: instruction dtype: string - name: preference dtype: int64 splits: - name: train num_bytes: 27875579 num_examples: 49383 - name: validation num_bytes: 1139961 num_examples: 2000 download_size: 15731882 dataset_size: 29015540 configs: - config_name: default data_files: - split: train path: data/train-* - split: validation path: data/validation-* ---

原始信息汇总

数据集信息

特征

• answers: 序列类型，字符串
• input: 数据类型，字符串
• instruction: 数据类型，字符串
• preference: 数据类型，int64

数据分割

• train:
  • 字节数: 27875579
  • 样本数: 49383
• validation:
  • 字节数: 1139961
  • 样本数: 2000

数据大小

• 下载大小: 15731882
• 数据集大小: 29015540

配置

• config_name: default
  • data_files:
    • train: data/train-*
    • validation: data/validation-*