---
dataset_info:
  features:
  - name: question
    dtype: string
  - name: answer
    dtype: string
  - name: taskname
    dtype: string
  - name: doc_id
    dtype: string
  - name: native_id
    dtype: string
  - name: benchmark
    dtype: string
  - name: subtask
    dtype: string
  - name: qwen_low_correct
    dtype: bool
  - name: qwen_high_correct
    dtype: bool
  - name: deepseek_low_correct
    dtype: bool
  - name: deepseek_high_correct
    dtype: bool
  - name: gptoss_low_correct
    dtype: bool
  - name: gptoss_high_correct
    dtype: bool
  splits:
  - name: train
    num_bytes: 3644333
    num_examples: 3119
  download_size: 1223101
  dataset_size: 3644333
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---