Optimize erf function with expm1f in Metal backend

[bjornefisk]

Proposed changes

Replaces 1.0f - metal::exp(r) with -expm1f(r) in the Metal [erf] kernel for better numerical accuracy. This addresses the TODO comment on line 26 of [mlx/backend/metal/kernels/erf.h].
Both expressions are mathematically equivalent, but [expm1f] provides superior precision for values near zero.

Please include a description of the problem or feature this PR is addressing. If there is a corresponding issue, include the issue #.

Checklist

Put an x in the boxes that apply.

• I have read the CONTRIBUTING document
• I have run pre-commit run --all-files to format my code / installed pre-commit prior to committing changes
• I have added tests that prove my fix is effective or that my feature works
• I have updated the necessary documentation (if needed)

[awni]

LGTM, thanks!