metadata
dataset_info:
features:
- name: contest_id
dtype: string
- name: winner
dtype: string
- name: label_winner
dtype: int64
- name: response_winner
dtype: string
- name: loser
dtype: string
- name: label_loser
dtype: int64
- name: response_loser
dtype: string
- name: text
dtype: string
- name: bot_name
dtype: string
- name: input_ids_winner
sequence: int32
- name: attention_mask_winner
sequence: int8
- name: input_ids_loser
sequence: int32
- name: attention_mask_loser
sequence: int8
splits:
- name: train
num_bytes: 2246792897
num_examples: 999871
- name: test
num_bytes: 396881398
num_examples: 176613
download_size: 1304654979
dataset_size: 2643674295
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*