config for parsing MATH

Hi! Currently debugging and updating lm_evals's implementation of `minerva_math`. Currently we basically use:

```python
gold = parse(normalize_final_answer(remove_boxed(last_boxed_only_string(doc["solution"]))))
answer = parse(raw_model_output)
```
and then `verify(gold, answer)`

however, this does have some edge cases for example for a gold (after removing boxed):

```python
parse("\\dfrac{9}{7}")
>>> []
```
but with boxed works correctly:

```python
parse("\\boxed{\\dfrac{9}{7}}")
>>> [9/7, '\\frac{9}{7}']
```

I was looking at how `lighteval` does it, and would [this](https://github.com/huggingface/lighteval/blob/b0edbc340485bd3581da5613a9183bd61aa4cce5/src/lighteval/metrics/metrics.py#L359-L361) generally work on most MATH tasks, or do y'll handle sub tasks differently? Should we also normalize before parsing?

Would appreciate any thoughts!