metadata
dataset_info:
features:
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: conversation_history
list:
- name: content
dtype: string
- name: role
dtype: string
- name: benchmark_name
dtype: string
- name: prompt_metadata
struct:
- name: rubrics
dtype: string
splits:
- name: train
num_bytes: 2181684
num_examples: 402
download_size: 1214600
dataset_size: 2181684
configs:
- config_name: default
data_files:
- split: train
path: data/train-*