metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: task_id
dtype: string
- name: input_ids_prompt
sequence: int64
- name: attention_mask_prompt
sequence: int64
splits:
- name: train
num_bytes: 4999207
num_examples: 1000
- name: test
num_bytes: 1001377
num_examples: 200
download_size: 101667
dataset_size: 6000584
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*