metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: responses
list: string
- name: train
dtype: string
- name: test
dtype: string
- name: source
dtype: string
- name: answer
dtype: string
- name: num_tokens
dtype: int64
- name: avg_hamming_distance
dtype: float64
- name: avg_percentage_match
dtype: float64
- name: avg_is_correct
dtype: float64
- name: avg_response_length
dtype: float64
- name: hamming_distances
sequence: int64
- name: percentage_matches
sequence: float64
- name: is_corrects
sequence: bool
- name: response_lengths
sequence: int64
- name: extracted_responses
sequence:
sequence:
sequence: int64
splits:
- name: test
num_bytes: 49394048
num_examples: 120
download_size: 17104020
dataset_size: 49394048
configs:
- config_name: default
data_files:
- split: test
path: data/test-*