Dialogue Reasoning on DIPLOMAT
168.2AIBC ScoreAutoMix + T
Evaluation results
| Method | Links | |
|---|---|---|
| AutoMix + TSLM=GPT-3.5, Router=Thresholding2023.10 | 168.2 | |
| AutoMix + PSLM=MISTRAL-7B, Router=POMDP-based2023.10 | 156.8 | |
| AutoMix + PSLM=GPT-3.5, Router=POMDP-based2023.10 | 151.2 | |
| AutoMix + TSLM=MISTRAL-7B, Router=Thresholding2023.10 | 149.7 | |
| HybridLLMSLM=MISTRAL-7B2023.10 | 67.1 | |
| AutoMix + PSLM=LLAMA2-13B, Router=POMDP-based2023.10 | 58.5 | |
| AutoMix + TSLM=LLAMA2-13B, Router=Thresholding2023.10 | 50.1 | |
| FrugalGPTSLM=GPT-3.52023.10 | 30.1 | |
| FrugalGPTSLM=MISTRAL-7B2023.10 | 16.8 | |
| HybridLLMSLM=GPT-3.52023.10 | 8.3 | |
| HybridLLMSLM=LLAMA2-13B2023.10 | 3.8 | |
| FrugalGPTSLM=LLAMA2-13B2023.10 | -7 |