metadata
dataset_info:
features:
- name: question_id
dtype: string
- name: problem
dtype: string
- name: solution
dtype: string
- name: grading_guidelines
dtype: string
- name: category
dtype: string
- name: level
dtype: string
- name: answer
dtype: string
- name: source
dtype: string
- name: model_solution
dtype: string
- name: reasoning_trace
dtype: string
- name: cost_run
dtype: string
- name: grade_cost
dtype: string
- name: schema_0
dtype: string
- name: grade
dtype: string
- name: score
dtype: float64
- name: contradiction
dtype: int64
- name: claim_lemma
dtype: int64
- name: direct_proof
dtype: int64
- name: substitution_specialization
dtype: int64
- name: wlog_symmetry
dtype: int64
- name: extremal_principle
dtype: int64
- name: constructive_proof
dtype: int64
- name: casework
dtype: int64
- name: forward_reference
dtype: int64
- name: polynomial_manipulation
dtype: int64
- name: induction
dtype: int64
- name: modular_arithmetic
dtype: int64
- name: boundedness_argument
dtype: int64
- name: pigeonhole
dtype: int64
- name: self_interruption
dtype: int64
- name: self_correction
dtype: int64
- name: error_acknowledgment
dtype: int64
- name: exploration
dtype: int64
- name: redo_restart
dtype: int64
- name: hedging
dtype: int64
- name: backtracking
dtype: int64
- name: structured_steps
dtype: int64
- name: verification
dtype: int64
splits:
- name: train
num_bytes: 29175751
num_examples: 960
download_size: 10948700
dataset_size: 29175751
configs:
- config_name: default
data_files:
- split: train
path: data/train-*