Response Generation on HuggingGPT Human Evaluation Set 130 diverse requests (test)
63.08Success RateHuggingGPT
Evaluation results
| Method | Links | |
|---|---|---|
| HuggingGPTLLM=GPT-3.52023.03 | 63.08 | |
| HuggingGPTLLM=Vicuna-13b2023.03 | 15.64 | |
| HuggingGPTLLM=Alpaca-13b2023.03 | 6.92 |