Clue Generation on COKE 1.0 (test)
37METEORCOLM
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| COLM2023.05 | 37 | 38.1 | 34 | 1.1 | 87.6 | |
| GPT-3.5 Turboshots=5-shot2023.05 | 29.3 | 26.6 | 25.9 | 0.8 | 87.2 | |
| GPT-4shots=5-shot2023.05 | 29.2 | 27 | 26.9 | 0.9 | 87.2 | |
| GPT-4shots=2-shot2023.05 | 28.6 | 26 | 24.9 | 0.8 | 87.1 | |
| GPT-3.5 Turboshots=2-shot2023.05 | 26.8 | 24.5 | 22.7 | 0.7 | 86.8 | |
| LLaMa-7Bshots=0-shot2023.05 | 25.1 | 18.3 | 11.5 | 0.3 | 84 | |
| LLaMa-7Bshots=2-shot2023.05 | 24.7 | 18.1 | 11.7 | 0.3 | 84.2 | |
| LLaMa-7Bshots=5-shot2023.05 | 24.6 | 17.7 | 11.9 | 0.3 | 84.4 | |
| Mistral-7Bshots=2-shot2023.05 | 23.8 | 17.8 | 11.6 | 0.3 | 83.2 | |
| Mistral-7Bshots=0-shot2023.05 | 23.3 | 17.2 | 11.7 | 0.3 | 82.9 | |
| Mistral-7Bshots=5-shot2023.05 | 22.4 | 18.4 | 11.7 | 0.4 | 84.1 | |
| GPT-3.5 Turboshots=0-shot2023.05 | 21.5 | 22.2 | 20 | 0.7 | 86.2 | |
| GPT-4shots=0-shot2023.05 | 21.1 | 19 | 18.3 | 0.5 | 85.9 |