Refactor: Improve clarity and robustness in route_tokens_to_experts

[danielquintas8]

What does this PR do?

This PR includes two small improvements for route_tokens_to_experts in MoE models:

• Improves Readability: Renames the function's input argument from hidden_states to router_logits. This more accurately reflects that the function receives the output of the gating network.
• Improves Robustness: Changes the softmax dimension from dim=1 to dim=-1. While functionally identical in the current implementation, dim=-1 is more explicit and ensures the logic remains correct if the tensor shape ever changes.

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

@ArthurZucker
@Cyrilvallez

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: dbrx, ernie4_5_moe, flex_olmo, hunyuan_v1_moe, jamba, olmoe, qwen2_moe, qwen3_moe, qwen3_next, qwen3_omni_moe