metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: sft_dpo
dtype: string
- name: base_dpo
dtype: string
- name: judgement_raw
sequence: string
- name: final_judgement
dtype: int64
splits:
- name: train
num_bytes: 407076
num_examples: 50
download_size: 141244
dataset_size: 407076
configs:
- config_name: default
data_files:
- split: train
path: data/train-*