LLM-rated generation quality on Dolly
4.1CorrectnessRole-aware LLM-Gen
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Role-aware LLM-GenBackbone=Qwen2.5 7B Instruct2025.07 | 4.1 | 3.7 | 4.7 | |
| Role-aware LLM-GenBackbone=Llama3.1 8B Instruct2025.07 | 4.1 | 3.7 | 4.7 | |
| Role-aware LLM-GenBackbone=Llama3.2 3B Instruct2025.07 | 4 | 3.6 | 4.7 | |
| Role-aware LLM-GenBackbone=Gemma3 4B Instruct2025.07 | 4 | 3.6 | 4.6 | |
| Role-aware LLM-GenBackbone=Qwen2.5 3B Instruct2025.07 | 3.9 | 3.5 | 4.6 | |
| Role-aware LLM-GenBackbone=Gemma 7B Instruct2025.07 | 3.9 | 3.5 | 4.5 |