metadata
dataset_info:
features:
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: rubrics
list: string
- name: metadata
struct:
- name: task_id
dtype: string
- name: context_id
dtype: string
- name: context_category
dtype: string
- name: sub_category
dtype: string
splits:
- name: train
num_bytes: 4068107
num_examples: 50
download_size: 4075751
dataset_size: 4068107
configs:
- config_name: default
data_files:
- split: train
path: data/train-*