| language: | |
| - en | |
| tags: | |
| - math | |
| - mathematics | |
| - reasoning | |
| - arithmetic | |
| - algebra | |
| - word-problems | |
| - logic | |
| - chain-of-thought | |
| - step-by-step | |
| - education | |
| - synthetic | |
| - llm | |
| - llm-training | |
| - pretraining | |
| - finetuning | |
| - Math | |
| - Logic | |
| - Algebra | |
| pretty_name: MathReasoning-2250000 | |
| size_categories: | |
| - 1M<n<10M | |
| license: mit | |
| # MathReasoning-2250000 | |
| ## A synthetic **math reasoning dataset** containing **2250000 examples**. | |
| ## Each row follows a strict reasoning-friendly structure: | |
| - `question` | |
| - `problem` | |
| - `how_to_solve` | |
| - `answer` | |
| This format is designed for: | |
| # - LLM reasoning | |
| # - chain-of-thought style supervision | |
| # - math finetuning | |
| # - educational tasks | |
| --- | |
| ## 📐 Problem Types | |
| # - Addition | |
| # - Subtraction | |
| # - Multiplication | |
| # - Division (integer-safe) | |
| --- | |
| ## 🧪 Recommended Splits | |
| # - Train: 80% | |
| # - Test: 20% | |
| --- | |
| ## Generated fully in-memory. No external data sources. | |