CenterforOpenScience: add ReplicatorBench benchmark

[masonmq]

From Center for Open Science:
1.Adds the ReplicatorBench benchmark implementation under hal/benchmarks/replicatorbench/.
2.Registers replicatorbench in hal/benchmark_manager.py.
3.Includes tasks and schemas/templates needed for evaluation.

Thank you!
ReplicatorBench [preprint]
Center for Open Science

[masonmq]

Hi HAL team @benediktstroebl @ksayash @schwartzadev :

This PR:

1. Adds the ReplicatorBench benchmark under hal/benchmarks/replicatorbench/
2. Registers replicatorbench in hal/benchmark_manager.py
3. Includes tasks + schemas/templates needed for evaluation.

Note: this benchmark submission is intentionally separated from the ReplicatorBench Agent submission. We are preparing the ReplicatorBench Agent as a separate PR so the benchmark review stays focused and easier to review.

Thank you!
ReplicatorBench [preprint]
Center for Open Science

[masonmq]

Hi HAL team @benediktstroebl , @ksayash , @schwartzadev :

Gentle follow-up on this PR. Our benchmark code is ready for review. We have validated that the our benchmark runs successfully using hal-eval in the harness.

We've also prepared our Agent as a separate PR so this benchmark review stays focused and easier to review.

Thank you!

[masonmq]

Hi HAL team :

Gentle follow-up on this PR again.

Our benchmark code is ready for review. We have validated that the our benchmark runs successfully using hal-eval in the harness.

We've also prepared our Agent as a separate PR so this benchmark review stays focused and easier to review.

Thank you!