metadata
dataset_info:
features:
- name: _id
dtype: string
- name: prompt
dtype: string
- name: chosen_completions
sequence: string
- name: rejected_completions
sequence: string
- name: chosen_rewards
sequence: float64
- name: chosen_rewards_raw
sequence:
sequence: float64
- name: rejected_rewards
sequence: float64
- name: rejected_rewards_raw
sequence:
sequence: float64
splits:
- name: train
num_bytes: 28964696
num_examples: 10795
download_size: 13940802
dataset_size: 28964696
configs:
- config_name: default
data_files:
- split: train
path: data/train-*