metadata
dataset_info:
features:
- name: instructions
dtype: string
- name: target_responses
dtype: string
- name: candidate_responses
dtype: string
- name: model_id
dtype: string
- name: model_sha
dtype: string
- name: similarity_scores
dtype: float64
- name: precision_scores
dtype: float64
- name: evaluators
dtype: string
- name: dates
dtype: string
splits:
- name: train
num_bytes: 147243
num_examples: 60
download_size: 46515
dataset_size: 147243
configs:
- config_name: default
data_files:
- split: train
path: data/train-*