metadata
dataset_info:
features:
- name: text
dtype: string
- name: source
struct:
- name: domain
dtype: string
- name: id
dtype: int64
- name: lm_label
dtype: string
- name: lm_name
dtype: string
- name: lm_q1_score
dtype: float64
- name: lm_q1q2_score
dtype: float64
- name: lm_q2_score
dtype: float64
- name: openwebmath_perplexity
dtype: float64
- name: openwebmath_score
dtype: float64
- name: tags
dtype: string
- name: url
dtype: string
splits:
- name: train
num_bytes: 15241780
num_examples: 100000
download_size: 5721938
dataset_size: 15241780
configs:
- config_name: default
data_files:
- split: train
path: data/train-*