metadata
dataset_info:
features:
- name: context
dtype: string
- name: question
dtype: string
- name: human_positive
dtype: string
- name: rag_positive
dtype: string
- name: hallucinated_span
dtype: string
- name: confidence
dtype: float64
splits:
- name: train
num_bytes: 1818726
num_examples: 2152
download_size: 500509
dataset_size: 1818726
configs:
- config_name: default
data_files:
- split: train
path: data/train-*