Empathetic Dialogue Generation on EMPATHETICDIALOGUES (test)
58.9Win RateCASE
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CASEComparison=vs. EmpDG, Aspect=Informativeness2022.08 | 58.9 | 31.6 | 0.5 | — | — | — | |
| CASEComparison=vs. CEM, Aspect=Informativeness2022.08 | 57.8 | 29.8 | 0.56 | — | — | — | |
| STRIDE-EDComparison Baseline=CAB, Evaluation Aspect=Relevance2026.04 | 56.6 | 21.4 | — | — | — | 22 | |
| CASEComparison=vs. CEM, Aspect=Empathy2022.08 | 53.2 | 34.6 | 0.47 | — | — | — | |
| STRIDE-EDComparison Baseline=CAB, Evaluation Aspect=Empathy2026.04 | 53.2 | 23.2 | — | — | — | 23.6 | |
| CASEComparison=vs. EmpDG, Aspect=Empathy2022.08 | 51.9 | 32.9 | 0.55 | — | — | — | |
| STRIDE-EDComparison Baseline=EmpDG, Evaluation Aspect=Empathy2026.04 | 51.6 | 18.5 | — | — | — | 29.9 | |
| CASEComparison=vs. KEMP, Aspect=Informativeness2022.08 | 51.1 | 34 | 0.53 | — | — | — | |
| STRIDE-EDComparison Baseline=EmpDG, Evaluation Aspect=Relevance2026.04 | 50.3 | 15.3 | — | — | — | 34.4 | |
| CASEComparison=vs. KEMP, Aspect=Empathy2022.08 | 50 | 34.4 | 0.53 | — | — | — | |
| CASEComparison=vs. EmpDG, Aspect=Coherence2022.08 | 48.1 | 39.2 | 0.54 | — | — | — | |
| Focused S1backbone=MIME, comparison=vs S02021.09 | 46.7 | 20 | — | — | — | 33.3 | |
| STRIDE-EDComparison Baseline=MOEL, Evaluation Aspect=Empathy2026.04 | 46.1 | 23.6 | — | — | — | 30.3 | |
| CASEComparison=vs. CEM, Aspect=Coherence2022.08 | 45.9 | 42.2 | 0.51 | — | — | — | |
| Focused S1backbone=Blender, comparison=vs S02021.09 | 44.6 | 37.4 | — | — | — | 18 | |
| CASEComparison=vs. KEMP, Aspect=Coherence2022.08 | 44.4 | 41.8 | 0.45 | — | — | — | |
| CEMComparison=vs. CASE, Aspect=Coherence2022.08 | 42.2 | 45.9 | 0.51 | — | — | — | |
| Focused S1backbone=DodecaTransformer, comparison=vs S02021.09 | 42.1 | 28.8 | — | — | — | 29.1 | |
| KEMPComparison=vs. CASE, Aspect=Coherence2022.08 | 41.8 | 44.4 | 0.45 | — | — | — | |
| STRIDE-EDComparison Baseline=MOEL, Evaluation Aspect=Relevance2026.04 | 39.8 | 22.4 | — | — | — | 37.8 | |
| EmpDGComparison=vs. CASE, Aspect=Coherence2022.08 | 39.2 | 48.1 | 0.54 | — | — | — | |
| STRIDE-EDComparison Baseline=EmpDG, Evaluation Aspect=Fluency2026.04 | 35.1 | 13.5 | — | — | — | 51.4 | |
| CEMComparison=vs. CASE, Aspect=Empathy2022.08 | 34.6 | 53.2 | 0.47 | — | — | — | |
| KEMPComparison=vs. CASE, Aspect=Empathy2022.08 | 34.4 | 50 | 0.53 | — | — | — | |
| KEMPComparison=vs. CASE, Aspect=Informativeness2022.08 | 34 | 51.1 | 0.53 | — | — | — | |
| STRIDE-EDComparison Baseline=MOEL, Evaluation Aspect=Fluency2026.04 | 33.5 | 15.6 | — | — | — | 50.9 | |
| EmpDGComparison=vs. CASE, Aspect=Empathy2022.08 | 32.9 | 51.9 | 0.55 | — | — | — | |
| EmpDGComparison=vs. CASE, Aspect=Informativeness2022.08 | 31.6 | 58.9 | 0.5 | — | — | — | |
| STRIDE-EDComparison Baseline=CAB, Evaluation Aspect=Fluency2026.04 | 31.5 | 11.6 | — | — | — | 56.9 | |
| CEMComparison=vs. CASE, Aspect=Informativeness2022.08 | 29.8 | 57.8 | 0.56 | — | — | — |