Chat Quality Evaluation on MT-Bench ShareGPT
6.12ScoreFedCoFiT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FedCoFiTFoundation Model=LLaMA22026.05 | 6.12 | 0.002 | |
| FLoRAFoundation Model=LLaMA22026.05 | 6.03 | 5.77 | |
| FlexLoRAFoundation Model=LLaMA22026.05 | 5.97 | 1.05 | |
| FedITFoundation Model=LLaMA22026.05 | 5.93 | 0.52 | |
| Centralized (LoRA)Foundation Model=LLaMA22026.05 | 5.86 | — | |
| FedCoFiTFoundation Model=LLaMA2026.05 | 3.45 | 0.002 | |
| FLoRAFoundation Model=LLaMA2026.05 | 3.41 | 6.875 | |
| FlexLoRAFoundation Model=LLaMA2026.05 | 3.36 | 1.25 | |
| Centralized (LoRA)Foundation Model=LLaMA2026.05 | 3.34 | — | |
| FedITFoundation Model=LLaMA2026.05 | 3.29 | 0.625 | |
| FLoRAFoundation Model=TinyLlama2026.05 | 2.81 | 1.98 | |
| FedCoFiTFoundation Model=TinyLlama2026.05 | 2.79 | 0.002 | |
| Centralized (LoRA)Foundation Model=TinyLlama2026.05 | 2.74 | — | |
| FlexLoRAFoundation Model=TinyLlama2026.05 | 2.73 | 0.36 | |
| FedITFoundation Model=TinyLlama2026.05 | 2.53 | 0.18 |