metadata
dataset_info:
features:
- name: id
dtype: string
- name: domain
dtype: string
- name: title
dtype: string
- name: physics_level
dtype: string
- name: problem_type
dtype: string
- name: natural_language_problem
dtype: string
- name: lean_problem_statement
dtype: string
- name: lean_answer_key
dtype: string
- name: solution_approach
dtype: string
- name: key_concepts
sequence: string
- name: has_error
dtype: bool
- name: error_info
struct:
- name: corrupted_token
dtype: string
- name: error_explanation
dtype: string
- name: error_subtype
dtype: string
- name: error_type
dtype: string
- name: original_token
dtype: string
- name: format
dtype: string
splits:
- name: train
num_bytes: 6404
num_examples: 5
- name: validation
num_bytes: 1624
num_examples: 1
- name: test
num_bytes: 2960
num_examples: 2
download_size: 42630
dataset_size: 10988
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*