metadata
dataset_info:
features:
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: chosen_score
dtype: int64
- name: rejected_score
dtype: int64
splits:
- name: train
num_bytes: 93679649
num_examples: 40482
- name: test
num_bytes: 12290336
num_examples: 4929
download_size: 53660570
dataset_size: 105969985
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*