metadata
dataset_info:
features:
- name: question
dtype: string
- name: answer
dtype: string
- name: question_source
dtype: string
- name: answer_source
dtype: string
- name: category
dtype: string
- name: ground_truth
dtype: string
- name: test_case
dtype: 'null'
- name: instruction_constrain
dtype: string
- name: pass_rate_r1
dtype: float64
- name: pass_rate_7b
dtype: float64
- name: pass_rate_1.5b
dtype: float64
- name: verify_score
dtype: float64
- name: ppl
dtype: float64
- name: model_name
dtype: string
splits:
- name: train
num_bytes: 1370071872.7857716
num_examples: 208609
download_size: 664751295
dataset_size: 1370071872.7857716
configs:
- config_name: default
data_files:
- split: train
path: data/train-*