metadata
configs:
- config_name: default
data_files:
- split: eval
path: data/eval-*
dataset_info:
features:
- name: instructions
dtype: string
- name: target_responses
dtype: string
- name: candidate_responses
dtype: string
- name: eval_prompts
dtype: string
- name: similarity_scores
dtype: int64
- name: precision_scores
dtype: int64
splits:
- name: eval
num_bytes: 70744
num_examples: 16
download_size: 62215
dataset_size: 70744