metadata
dataset_info:
features:
- name: label
dtype: float64
- name: category
dtype: int64
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: input_ids_chosen
sequence: int64
- name: attention_mask_chosen
sequence: int64
- name: input_ids_rejected
sequence: int64
- name: attention_mask_rejected
sequence: int64
- name: length
dtype: int64
splits:
- name: train
num_bytes: 11460330
num_examples: 14210
- name: test
num_bytes: 1195070
num_examples: 1579
download_size: 2173041
dataset_size: 12655400
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*