#4: WIP DRAFT accuracy eval and evaluator; add GPQA for testing

[nvzhihanj]

Address part 1 of #4

Added:

• Evaluator for GPQA and AIME (refered to GPT-OSS)
• Supports for repeat and pass@k (note that repeats will have N-length responses but only 1-length ground truth for ease of logics. The current assumption is that we don't repeat ground truths or input pickle).
• Interface for eval and eval-results CLI
• eval needs integration to the benchmark (and requires a deterministic SampleOrder), and changes to accommodate repeats

TODOs (in this PR):

• Add integration tests for eval-results (a sample
• add unit tests for evaluator components

TODOs (in next PR):

• Hook up eval to benchmarks
• Add GPT-OSS chat template and system prompt integration for eval
• Add eval example (using small/full GPQA dataset on GPT-OSS)

What does this PR do?

Type of change

• Bug fix
• New feature
• Documentation update
• Refactor/cleanup

Related issues

Testing

• Tests added/updated
• All tests pass locally
• Manual testing completed

Checklist

• Code follows project style
• Pre-commit hooks pass
• Documentation updated (if needed)

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[nv-alicheng]

Do we still need this feature? It looks like it's only used for debugging / testing