Added additional bridge analysis tools

[jlarson4]

Description

• W_E identity: same tensor object as embed.W_E (verified via is)
• W_U/b_U equality: same values as unembed.W_U/unembed.b_U (view, so torch.equal not is)
• W_U matches HookedTransformer: max diff < 1e-4
• tokens_to_residual_directions: result equals W_U[:, token] exactly (torch.equal), matches HT
• accumulated_bias: layer 0 is zeros, layer 1+ is non-zero, matches HT for all layers including mlp_input flag
• all_composition_scores: correct shape, upper-triangular masking enforced, non-zero above diagonal, all 3 modes work, invalid mode raises
• all_head_labels: correct count, format, matches HT exactly

Type of change

Please delete options that are not relevant.

• New feature (non-breaking change which adds functionality)

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