Conversational AI on Combined PA platforms (SIGESON and SIDFORS) (test)
61.14ROUGE-1cerbero-7b
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| cerbero-7bLearning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 61.14 | 53.72 | 59.09 | 44.3 | 62.99 | |
| cerbero-7b-openchatLearning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 61.1 | 55.77 | 59.44 | 45.02 | 63.94 | |
| zefiro-7b-beta-ITA-v0.1Learning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 58.55 | 50.97 | 56.3 | 39.43 | 61.57 | |
| Loquace-7B-MistralLearning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 53.1 | 46.81 | 51.68 | 37.63 | 54.75 | |
| Mistral-7B-Instruct-v0.3Learning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 53.02 | 43.26 | 49.91 | 34.33 | 54.27 | |
| Mistral-7B-v0.1Learning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 50.3 | 42.38 | 47.63 | 34.53 | 53.71 | |
| maestrale-chat-v0.4-betaLearning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 49.25 | 40.9 | 46.82 | 32.26 | 50.05 | |
| Llama-3.1-8B-InstructLearning Setup=Federated Learning, Aggregation Strategy=Global, Evaluation Stage=Final Communication Round2026.05 | 45.86 | 38.45 | 44.83 | 29.32 | 50.83 |