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
- name: c
dtype: float64
splits:
- name: train
num_bytes: 114879635
num_examples: 81157
- name: test
num_bytes: 14745784
num_examples: 10722
download_size: 14908033
dataset_size: 129625419
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*