metadata
dataset_info:
features:
- name: id
dtype: int64
- name: model_a
dtype: string
- name: model_b
dtype: string
- name: prompt
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: winner_model_a
dtype: int64
- name: winner_model_b
dtype: int64
- name: winner_tie
dtype: int64
- name: sw_ranking
dtype: float64
- name: matrix_factorization
dtype: float64
- name: bert
dtype: float32
splits:
- name: train
num_bytes: 185877048
num_examples: 57477
download_size: 102826088
dataset_size: 185877048
configs:
- config_name: default
data_files:
- split: train
path: data/train-*