Add a Simplified Chinese localization of Transformer Explainer

[Jian1202]

Closes #57.

Summary

This PR implements the Simplified Chinese (zh-CN) localization proposed in #57.
It aims to make Transformer Explainer more accessible to Chinese-speaking learners
while preserving the project’s interactive learning experience.

Included changes

• Translated the interface, tooltips, tutorials, and explanatory materials.
• Added localized feedback for token counts, model loading, cache state, and failures.
• Improved responsive behavior for the localized experience.
• Corrected several GPT-2 teaching explanations and examples.
• Updated the localized README, including the scope and known limitations of the Chinese version.
• Updated project dependencies and static-build configuration so the localized fork can be built and deployed reliably.

Scope and limitations

• The underlying checkpoint remains English GPT-2 Small; this PR does not introduce a Chinese language model.
• Chinese text can be tokenized, but byte-level token fragments may display as � in per-token visualizations. This is a display limitation rather than a page encoding issue.
• Some GitHub Pages paths and localized-site links are specific to the Chinese fork. If maintainers prefer, I can separate those deployment-specific changes from the localization work.

Validation

• npm ci
• npm run build
• npm run check:localization
• Manual checks of generation, sampling controls, visualizations, tutorials, and responsive layouts