File size: 2,166 Bytes
a7afdd1 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 | ---
language:
- en
pretty_name: IO-Bench
license: cc-by-nd-4.0
size_categories:
- n<1K
task_categories:
- question-answering
- text-generation
tags:
- economics
- mathematical-reasoning
- symbolic-reasoning
- benchmark
- responsible-ai
configs:
- config_name: default
data_files:
- split: test
path: data/test.jsonl
---
# IO-Bench
IO-Bench is a 155-example evaluation dataset for mathematical economics reasoning. Each record contains a standalone economics question, a reference answer, a machine-comparable answer field, symbolic answer metadata where applicable, and review status metadata.
Unless otherwise noted, the dataset materials in this repository are licensed under the Creative Commons Attribution-NoDerivatives 4.0 International License (CC BY-ND 4.0). See `LICENSE` for details.
## Dataset Structure
- Main HF split: `data/test.jsonl`
- Chapter copies: `data/by_chapter/*.jsonl`
- Problem-type labels: included in each JSONL record as `problem_type_label`
- Rows: 155
- Splits: {'chapter_1': 21, 'chapter_2': 10, 'chapter_3': 7, 'chapter_4': 9, 'chapter_5': 12, 'chapter_6': 17, 'chapter_7': 5, 'chapter_8': 17, 'chapter_9': 9, 'chapter_10': 15, 'chapter_11': 29, 'chapter_intro': 4}
## Fields
- `id`: Unique example id.
- `chapter`, `split`, `problem_number`, `sub_id`: Source location metadata after anonymization.
- `question`: Standalone benchmark question.
- `question_type`: Question format, such as `value` or `judge`.
- `problem_type_label`: Problem-type category used for analysis.
- `answer_kind`: Answer representation category.
- `comparison_mode`: Evaluation mode, such as `sympy`, `exact`, or `json`.
- `reference_answer`: Reference final answer.
- `reference_answer_sympy`: SymPy-compatible answer when applicable.
- `reference_answer_json`: JSON answer when applicable.
- `reference_answer_json_modes`: Per-field comparison modes for JSON answers when applicable.
- `answer_for_compare`: Canonical comparison target used by the evaluator.
- `reference_reasoning`: Reference solution rationale.
- `allowed_symbols`, `symbol_definitions`: Symbol contract for symbolic answers.
- `review_status`: Anonymized review status.
|