metadata
dataset_info:
features:
- name: problem
dtype: string
- name: source
dtype: string
- name: id
sequence: int64
- name: reasoning
dtype: string
- name: deepseek_solution
dtype: string
- name: __original_row_idx
dtype: int64
- name: final_reasoning_trace
sequence: string
- name: correct_majority_indices
sequence: int64
- name: _judge_reasoning
dtype: string
- name: _majority_responses
sequence: string
- name: verified_final_reasoning_trace
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 2620050
num_examples: 3
download_size: 511300
dataset_size: 2620050
configs:
- config_name: default
data_files:
- split: train
path: data/train-*