metadata
dataset_info:
features:
- name: prompt_id
dtype: string
- name: prompt
dtype: string
- name: responses
sequence: string
- name: scores
sequence: int64
- name: p_weight
dtype: float64
- name: p_weight_pairwise
dtype: float64
splits:
- name: train
num_bytes: 133944243
num_examples: 96403
- name: test
num_bytes: 14269695
num_examples: 9338
download_size: 15732284
dataset_size: 148213938
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*