Approximating Human-Like Few-shot Learning with GPT-based Compression

Re-assign community