Refactor sae classes

[curt-tigges]

Description

Working implementations of StandardSAE and StandardTrainingSAE, along with tests to compare it to old SAE and TrainingSAE implementations. References and folder structure also updated.

Significant change: The old TrainingSAE did not reshape hook_z activations, while the inference SAE did. The former assumed that the ActivationStore would reshape the activations. I suggest we move this functionality to the new training SAEs instead. (Even in cases where caching is used for activations, only ~4% of compute is spent on reshaping--seems worthwhile if it makes the two SAE classes work the same).

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

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 tests pass locally with my changes
• I have not rewritten tests relating to key interfaces which would affect backward compatibility

You have tested formatting, typing and tests

• I have run make check-ci to check format and linting. (you can run make format to format code if needed.)