Diversified Sampling Improves Scaling LLM inference

Open in new window