[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/corpus/open/math-scripts/)

[home](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [corpus](https://md.mova.elamur.ai/free/en/docs/corpus/) / [open](https://md.mova.elamur.ai/free/en/docs/corpus/open/) / Math world scripts

# Math world scripts

2,680 text word problems with a script that tells, step by step, what the problem's world is: who has what, what changes, what is asked. One JSON object per line: `id`, `set` (`svamp-train` or `gsm8k-train`), `question`, `gold` (answer), `script`.

Example script:

```
@1 set philip.orange = 87                          k=given
@1 set philip.banana = 290                         k=given
@2 set philip.banana_group = 2                     k=given
@3 set philip.banana_per_group = philip.banana / philip.banana_group   k=unit
@3 ask philip.banana_per_group
```

`@N` is the sentence the line comes from; `k=` is the kind of step.

**Sources:** problems from the training splits of [SVAMP](https://github.com/arkilpatel/SVAMP) (Patel et al. 2021, MIT) and [GSM8K](https://github.com/openai/grade-school-math) (Cobbe et al. 2021, MIT).

**How made:** a large language model wrote a script for each problem; Mova ran it with its exact arithmetic core and kept only scripts whose result equals the gold answer.

**Sources:** the problems are from SVAMP and GSM8K (MIT); the scripts are Mova's.

**How Mova uses it:** training data for Mova's own world writer in `math` (`qworld`): it learns to label each number with its role in the world, so that the judge of the step-by-step solver can check a candidate against the world.

- [← previous Dictionary](https://md.mova.elamur.ai/free/en/docs/corpus/open/dictionary/)
- [next → Treebanks](https://md.mova.elamur.ai/free/en/docs/corpus/open/treebanks/)
- [↑ back to Mova corpus 0.4](https://md.mova.elamur.ai/free/en/docs/corpus/open/)