0

Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models

The Beyond the Imitation Game Benchmark (BIG-bench) is a collaborative benchmark intended to probe large language models and extrapolate their future capabilities.