Decision making on vignette-based (inference)
48.3Vignette ScoreMeTHanol (8B)
Evaluation Results
| Method | Links | |
|---|---|---|
| MeTHanol (8B)Model size=similar size, Zero-shot=true2024.09 | 48.3 | |
| GPT-4Model size=much larger size, Zero-shot=true2024.09 | 46.9 | |
| Mistral-7B-InstructModel size=similar size, Zero-shot=true2024.09 | 40.2 | |
| GPT-3Model size=much larger size, Zero-shot=true2024.09 | 37.5 | |
| Llama3-8B-InstructModel size=similar size, Zero-shot=true2024.09 | 23.8 | |
| Quiet-STaR (7B)Model size=similar size, Zero-shot=true2024.09 | 11.1 |