metadata
dataset_info:
features:
- name: text
dtype: string
- name: category
dtype: string
- name: url
dtype: string
- name: title
dtype: string
- name: __index_level_0__
dtype: int64
- name: problem
dtype: string
- name: __original_row_idx
dtype: int64
- name: reasoning
dtype: string
- name: deepseek_solution
dtype: string
- name: final_reasoning_trace
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: discipline
dtype: string
- name: expert
dtype: string
- name: num_topics
dtype: int64
- name: num_subtopics
dtype: int64
- name: num_questions
dtype: int64
- name: topic
dtype: string
- name: subtopic
dtype: string
- name: score
dtype: int64
- name: year
dtype: int64
- name: id
dtype: string
splits:
- name: train
num_bytes: 1585354663.3333333
num_examples: 50000
download_size: 774907484
dataset_size: 1585354663.3333333
configs:
- config_name: default
data_files:
- split: train
path: data/train-*