Benchmarking of prompt2model on composite benchmarks

[neubig]

Currently, we have benchmarked prompt2model extensively on three tasks (as detailed in our preprint).

But it would be much cooler if we could benchmark it on a bunch of tasks included in composite benchmarks. Examples of this include:

1. BIG-Bench
2. OpenAI evals
3. Eleuther Evaluation Harness

In order to do this, we'll have to

• [ ] create code that interfaces with each of these benchmarks
• [ ] run experiments
• [ ] probably finish https://github.com/neulab/prompt2model/issues/285