metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: metadata
struct:
- name: T_max
dtype: int64
- name: numbers
sequence:
sequence: int64
- name: target
sequence: int64
splits:
- name: train
num_bytes: 101390647
num_examples: 100000
- name: eval
num_bytes: 101407
num_examples: 100
download_size: 34246054
dataset_size: 101492054
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*