A Judge-free LLM Open-ended Generation Benchmark Based on the Distributional Hypothesis

Open in new window