metadata
dataset_info:
features:
- name: example_tags
sequence: string
- name: ideal_completions_data
struct:
- name: ideal_completion
dtype: string
- name: ideal_completions_group
dtype: string
- name: ideal_completions_ref_completions
sequence: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: prompt_id
dtype: string
- name: rubrics
list:
- name: criterion
dtype: string
- name: points
dtype: int64
- name: tags
sequence: string
- name: canary
dtype: string
splits:
- name: train
num_bytes: 55082653
num_examples: 5000
download_size: 25083470
dataset_size: 55082653
configs:
- config_name: default
data_files:
- split: train
path: data/train-*