metadata
dataset_info:
features:
- name: problem
dtype: string
- name: answer
dtype: string
- name: source
dtype: string
- name: domain
sequence: string
- name: llama8b_solve_rate
dtype: float64
- name: better_answer
dtype: string
splits:
- name: train
num_bytes: 65717832
num_examples: 198977
download_size: 28225453
dataset_size: 65717832
configs:
- config_name: default
data_files:
- split: train
path: data/train-*