Benchmarking: Centralize results layout #1324

Is your feature request related to a problem? Please describe.
GitHub as the single source of truth: Benchmark results currently live in two places: the Netlify page and Drive links/docs. We’ll bring everything together in the GitHub repo and use that as the single source of truth going forward, so there’s no need to look across multiple Drive links for numbers.

Describe the solution you'd like
Use the eval-run layout for GitHub Pages: Update GitHub Pages to follow this layout: https://rococo-cat-a45cd7.netlify.app/ Benchmark results for each validator will be available in one place, making it easier to scan and compare results without opening individual Drive links.

Limit benchmarks to validators exposed in Glific: Since the guardrails repo also needs cleanup, we’ll keep only the validators that are actually exposed in Glific and will be used in future. Validators we don’t plan to use will be removed instead of maintained.

Keep the PII dataset format consistent: We’ll update the new PII dataset to follow the same format as the other validator benchmarks, so the results are consistent across the page.