Added warnings for users attempting to use MPS with Torch 2.8

[jlarson4]

Description

Added an explicit warning when MPS users attempt to run TransformerLens with Torch 2.8.0, it has a known MPS bug that produces silently incorrect results (torch.nn.functional.linear on non-contiguous tensors). This corrupts generate() output and attention computations.

I did not do anything directly to processing to resolve this on our end in any way, this is an upstream issue & has been resolved as of Torch 2.9.0. This warning will alert anyone running on older dependencies to upgrade.

Fixes #1062

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have not rewritten tests relating to key interfaces which would affect backward compatibility