Datasets:

CZWin32768
/

multi-solution-countdown

dataset_info:
  features:
    - name: index
      dtype: int64
    - name: target
      dtype: int64
    - name: available_numbers
      sequence: int64
    - name: solutions
      sequence: string
  splits:
    - name: train
      num_bytes: 4524413
      num_examples: 22500
    - name: test
      num_bytes: 86631
      num_examples: 400
  download_size: 1757059
  dataset_size: 4611044
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*

Multi-Solution Countdown Dataset

This dataset is from the paper The Era of Agentic Organization: Learning to Organize with Language Models.

Dataset Description

The Multi-Solution Countdown dataset contains mathematical reasoning problems where the goal is to reach a target number using a set of available numbers and basic arithmetic operations (+, -, *, /). Each problem has multiple valid solutions.

Dataset Structure

Split	Examples
Train	22,500
Test	400

Features

index: Integer identifier
target: Target number to reach
available_numbers: List of numbers that can be used
solutions: List of valid mathematical expressions

Example

{
  "index": 1,
  "target": 655,
  "available_numbers": [8, 9, 26, 43, 47, 60, 68, 69, 70, 78, 82, 87],
  "solutions": ["((26-78)+((68+87)+(8*69)))", "(69-(70-(8*82)))", "(43+(68*9))", "((47+68)+(60*9))"]
}

Usage

from datasets import load_dataset

dataset = load_dataset("CZWin32768/multi-solution-countdown")

Citation

@article{chi2025asyncthink,
  title={The Era of Agentic Organization: Learning to Organize with Language Models},
  author={Chi, Zewen and Dong, Li and Dong, Qingxiu and Hao, Yaru and Wu, Xun and Huang, Shaohan and Wei, Furu},
  journal={arXiv preprint arXiv:2510.26658},
  year={2025}
}