metadata
dataset_info:
features:
- name: chosen
dtype: int64
- name: rejected
dtype: int64
- name: chosen_rank
dtype: int64
- name: rejected_rank
dtype: int64
- name: top_level_parent
dtype: int64
- name: split
dtype: large_string
- name: chosen_prompt
dtype: large_string
- name: rejected_prompt
dtype: large_string
splits:
- name: train
num_bytes: 115409826
num_examples: 30000
- name: validation
num_bytes: 1964047
num_examples: 500
- name: test
num_bytes: 3860648
num_examples: 1000
download_size: 68080869
dataset_size: 121234521
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*