metadata
dataset_info:
features:
- name: id
dtype: string
- name: context
dtype: string
- name: prompt
dtype: string
- name: response
dtype: string
- name: predict_label
dtype: float64
- name: top_k_context
dtype: string
splits:
- name: train
num_bytes: 4880537
num_examples: 2000
download_size: 2642327
dataset_size: 4880537
configs:
- config_name: default
data_files:
- split: train
path: data/train-*