BoCoEL

Python tool that speeds up LLM benchmarking by using Bayesian optimization to choose a small, representative subset of evaluation data.