ResearchBenchmarksModel Selection on HuggingGPT Human Evaluation Set 130 diverse requests (test)Model Selection on HuggingGPT Human Evaluation Set 130 diverse requests (test)Follow93.89Passing RateHuggingGPT89.195591.5427593.8996.23725Mar 30, 2023Evaluation resultsMethodMethodLinksPassing RateRationalityHuggingGPTLLM=GPT-3.5LLM=GPT-3.52023.0393.8981.29