Controversy Controllability Evaluation on Controversial Topics (test)
95.8Positive ScoreDEBATUNE-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DEBATUNE-7BParameters=7B2024.02 | 95.8 | 97.9 | 96.9 | |
| DEBATUNE-13BParameters=13B2024.02 | 95 | 94.6 | 94.8 | |
| Zephyr 7B BetaParameters=7B, Version=Beta2024.02 | 94.2 | 73.3 | 83.8 | |
| Alpaca 7BParameters=7B2024.02 | 93.8 | 88.3 | 91 | |
| Vicuna 7B v1.5Parameters=7B, Version=v1.52024.02 | 90 | 79.6 | 84.8 | |
| Zephyr 7B AlphaParameters=7B, Version=Alpha2024.02 | 87.9 | 71.3 | 79.6 | |
| Vicuna 13B v1.5Parameters=13B, Version=v1.52024.02 | 86.7 | 85.8 | 86.3 | |
| WizardLM 7BParameters=7B2024.02 | 83.3 | 70.4 | 76.8 | |
| WizardLM 13B V1.2Parameters=13B, Version=V1.22024.02 | 80 | 70.8 | 75.4 | |
| LLaMA2 Chat 13BParameters=13B2024.02 | 33.8 | 31.7 | 32.7 | |
| LLaMA2 Chat 7BParameters=7B2024.02 | 19.6 | 42.9 | 31.3 |