metadata
dataset_info:
features:
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: response
dtype: string
- name: reward_model
struct:
- name: ground_truth
dtype: string
- name: style
dtype: string
- name: ability
dtype: string
- name: extra_info
struct:
- name: index
dtype: int64
- name: split
dtype: string
- name: cross_domain
dtype: bool
- name: entities
sequence: string
- name: papers
sequence: string
- name: task_type
dtype: string
splits:
- name: train
num_bytes: 655120
num_examples: 636
- name: validation
num_bytes: 163886
num_examples: 160
- name: test
num_bytes: 194787
num_examples: 100
download_size: 219933
dataset_size: 1013793
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*