metadata
dataset_info:
features:
- name: traces
sequence: string
- name: scores
sequence: int64
- name: messages
list:
list:
- name: content
dtype: string
- name: role
dtype: string
- name: correct
dtype: bool
splits:
- name: train
num_bytes: 8060
num_examples: 1
download_size: 7213
dataset_size: 8060
configs:
- config_name: default
data_files:
- split: train
path: data/train-*