metadata
dataset_info:
features:
- name: item_id
dtype: string
- name: question
dtype: string
- name: answer_value
dtype: string
- name: prediction
dtype: string
- name: extracted_code
dtype: string
- name: predicted_answer
dtype: string
- name: correctness
dtype: bool
- name: correctness_from_feedback
dtype: bool
- name: system
dtype: string
- name: user
dtype: string
- name: assistant
dtype: string
splits:
- name: train
num_bytes: 53730158
num_examples: 13152
download_size: 15128690
dataset_size: 53730158
configs:
- config_name: default
data_files:
- split: train
path: data/train-*