Create TinyBenchmarks for Too Large Benchmarks

[Malikeh97]

• Create TinyBenchmarks for Too Large Benchmarks

• Put the dataset on R3 HF Space

• Push the data creation codes to the repo

• Integrate the benchmark with lm-eval R3 fork