metadata
dataset_info:
features:
- name: text
dtype: string
- name: label
dtype: string
- name: condition
dtype: string
- name: prompt
dtype: string
- name: rejected
dtype: string
- name: chosen
dtype: string
- name: input_ids_chosen
sequence: int64
- name: attention_mask_chosen
sequence: int64
- name: input_ids_rejected
sequence: int64
- name: attention_mask_rejected
sequence: int64
splits:
- name: train
num_bytes: 2072114.0594059406
num_examples: 80
- name: test
num_bytes: 543929.9405940594
num_examples: 21
download_size: 474198
dataset_size: 2616044
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*