Repository for examining humaneval and similar benchmarks.

Testing the differences in outputs on boolq for llama and gemma