Adding ruff, running pre-commit hooks, small fixes and documentation (#1303) This doesn't contribute an Eval but slightly improves the developer experience for contributors.

Make the torch dep optional (#1524) `torch` was added in https://github.com/openai/evals/pull/1496, but it's very heavy and only required for one eval. Let's move it to an optional-dependency

Use the API key for testing evals in CI (#1443) Passes in the API key for testing new evals via CI checks (note: only works for PRs made from a branch within the repo)