metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: target
dtype: string
- name: chunk_prompt
dtype: bool
- name: kind
dtype: string
- name: prob
dtype: float64
- name: path
dtype: string
- name: quality_prob
dtype: float64
- name: learning_prob
dtype: float64
- name: filename
dtype: string
splits:
- name: train
num_bytes: 7581106764.450301
num_examples: 870654
download_size: 3062286174
dataset_size: 7581106764.450301
configs:
- config_name: default
data_files:
- split: train
path: data/train-*