metadata
dataset_info:
features:
- name: pos_input_ids
sequence: int64
- name: neg_input_ids
sequence: int64
- name: pos_reward
dtype: float64
- name: neg_reward
dtype: float64
splits:
- name: train
num_bytes: 63292832
num_examples: 10000
download_size: 13950231
dataset_size: 63292832
configs:
- config_name: default
data_files:
- split: train
path: data/train-*