Dialogue Quality Assessment on ECG-MTD
3.98NaturalnessQwen3 32B (ECG-Agent)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3 32B (ECG-Agent)Evaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II, Base Model=Qwen3 32B2026.01 | 3.98 | 4.41 | |
| Gemini-2.5-FlashEvaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II2026.01 | 3.97 | 4.5 | |
| Llama 3.2 3B (ECG-Agent)Evaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II, Base Model=Llama 3.2 3B2026.01 | 3.96 | 4.37 | |
| Llama 3.2 1B (ECG-Agent)Evaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II, Base Model=Llama 3.2 1B2026.01 | 3.95 | 4.32 | |
| Llama 3.1 8B (ECG-Agent)Evaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II, Base Model=Llama 3.1 8B2026.01 | 3.95 | 4.38 | |
| PULSEEvaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II2026.01 | 3.67 | 3.57 | |
| GEMEvaluator=Gemini-2.5-Pro, Input Configuration=Average of 12-lead, Lead I, Lead II2026.01 | 3.63 | 3.51 |