Dialogue Response Generation on Topical-Chat Global
98.5UndGPT-4o mini
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPT-4o mini2025.02 | 98.5 | 83.3 | 82.6 | 76.7 | 87.7 | 85.8 | 80.5 | |
| GPT-4o2025.02 | 97.8 | 81.5 | 87.5 | 79.2 | 88.8 | 87 | 80.7 | |
| Qwen-2.5-72BParameters=72B2025.02 | 97.8 | 91 | 86.2 | 88.6 | 90.4 | 90.8 | 86.9 | |
| GPT-4 Turbo2025.02 | 97.2 | 77.4 | 84.1 | 84.9 | 88.9 | 86.5 | 79.5 | |
| Gemma-2-27BParameters=27B2025.02 | 96.6 | 72.4 | 75 | 75.1 | 84.2 | 80.6 | 78 | |
| Phi-4-14BParameters=14B2025.02 | 96.3 | 82.3 | 83.7 | 88.7 | 90.1 | 88.2 | 81.7 | |
| DeepSeek-V32025.02 | 94.7 | 76.9 | 81.9 | 77.1 | 89.4 | 84 | 77.9 | |
| CompassJudger-32BParameters=32B2025.02 | 94.6 | 81.8 | 81.9 | 92.8 | 83.5 | 86.9 | 86.2 | |
| Themis-8BParameters=8B2025.02 | 94.1 | 85.9 | 80.5 | 84.3 | 72.9 | 83.5 | 84 | |
| Llama-3.1-70BParameters=70B2025.02 | 91.4 | 80 | 78.9 | 76.6 | 81 | 81.6 | 76.8 | |
| GPT-3.5 Turbo2025.02 | 90.5 | 77.2 | 69.9 | 82.9 | 84.4 | 81 | 75.1 | |
| Prometheus-2-8x7BParameters=8x7B2025.02 | 84.9 | 74 | 79.3 | 87.8 | 82.7 | 81.8 | 79 | |
| Prometheus-2-7BParameters=7B2025.02 | 83.6 | 68.9 | 77.4 | 75.1 | 83.9 | 77.8 | 74.6 | |
| CRITIQUELLM-6BParameters=6B2025.02 | 76.9 | 57.7 | 60.2 | 67 | 77.6 | 67.9 | 69.3 | |
| Prometheus-13BParameters=13B2025.02 | 69.2 | 56.9 | 55.7 | 63.8 | 73.6 | 63.9 | 63 | |
| Auto-J-13BParameters=13B2025.02 | 61.4 | 60.2 | 58.9 | 66.6 | 61.2 | 61.7 | 61.3 |