Emotional Support Conversation on ESConv and ExTES (val)
68.5Win RateAFlow
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| AFlowOpponent=EmoDx, Evaluator=GPT-5.2, Metric Dimension=Suggest2026.02 | 68.5 | — | — | — | 68.5 | — | 18.5 | 13 | |
| AFlowOpponent=ESCoT, Evaluator=GPT-5.2, Metric Dimension=Comfort2026.02 | 65.2 | — | — | 65.2 | — | — | 20.5 | 14.3 | |
| AFlowOpponent=EmoDx, Evaluator=Human Evaluation, Metric Dimension=Suggest2026.02 | 64.2 | — | — | — | 64.2 | — | 21.5 | 14.3 | |
| AFlowOpponent=ESCoT, Evaluator=GPT-5.2, Metric Dimension=Overall2026.02 | 62.5 | — | — | — | — | 62.5 | 22.5 | 15 | |
| AFlowOpponent=ESCoT, Evaluator=Human Evaluation, Metric Dimension=Comfort2026.02 | 60.5 | — | — | 60.5 | — | — | 24 | 15.5 | |
| AFlowOpponent=ESCoT, Evaluator=Human Evaluation, Metric Dimension=Overall2026.02 | 58.8 | — | — | — | — | 58.8 | 25.2 | 16 | |
| AFlowOpponent=ESCoT, Evaluator=GPT-5.2, Metric Dimension=Identification2026.02 | 58.5 | — | 58.5 | — | — | — | 28.2 | 13.3 | |
| AFlowOpponent=EmoDx, Evaluator=GPT-5.2, Metric Dimension=Overall2026.02 | 58.2 | — | — | — | — | 58.2 | 25.5 | 16.3 | |
| AFlowOpponent=ESCoT, Evaluator=GPT-5.2, Metric Dimension=Suggest2026.02 | 55.4 | — | — | — | 55.4 | — | 26.8 | 17.8 | |
| AFlowOpponent=ESCoT, Evaluator=Human Evaluation, Metric Dimension=Identification2026.02 | 55 | — | 55 | — | — | — | 30.5 | 14.5 | |
| AFlowOpponent=EmoDx, Evaluator=Human Evaluation, Metric Dimension=Overall2026.02 | 54.5 | — | — | — | — | 54.5 | 28.5 | 17 | |
| AFlowOpponent=ESCoT, Evaluator=Human Evaluation, Metric Dimension=Suggest2026.02 | 52.5 | — | — | — | 52.5 | — | 29 | 18.5 | |
| AFlowOpponent=GPT-4o, Evaluator=GPT-5.2, Metric Dimension=Suggest2026.02 | 52.5 | — | — | — | 52.5 | — | 32.5 | 15 | |
| AFlowOpponent=Claude-3.5, Evaluator=GPT-5.2, Metric Dimension=Suggest2026.02 | 50.8 | — | — | — | 50.8 | — | 33.2 | 16 | |
| AFlowOpponent=EmoDx, Evaluator=GPT-5.2, Metric Dimension=Identification2026.02 | 50.5 | — | 50.5 | — | — | — | 35.5 | 14 | |
| AFlowOpponent=GPT-4o, Evaluator=GPT-5.2, Metric Dimension=Overall2026.02 | 50.5 | — | — | — | — | 50.5 | 33.5 | 16 | |
| AFlowOpponent=GPT-4o, Evaluator=Human Evaluation, Metric Dimension=Suggest2026.02 | 49.2 | — | — | — | 49.2 | — | 35.5 | 15.3 | |
| AFlowOpponent=Claude-3.5, Evaluator=GPT-5.2, Metric Dimension=Overall2026.02 | 48.8 | — | — | — | — | 48.8 | 36.2 | 15 | |
| AFlowOpponent=GPT-4o, Evaluator=GPT-5.2, Metric Dimension=Identification2026.02 | 48.5 | — | 48.5 | — | — | — | 38.5 | 13 | |
| AFlowOpponent=GPT-4o, Evaluator=Human Evaluation, Metric Dimension=Overall2026.02 | 48.5 | — | — | — | — | 48.5 | 35.2 | 16.3 | |
| AFlowOpponent=Claude-3.5, Evaluator=Human Evaluation, Metric Dimension=Suggest2026.02 | 48.5 | — | — | — | 48.5 | — | 35 | 16.5 | |
| AFlowOpponent=GPT-4o, Evaluator=GPT-5.2, Metric Dimension=Comfort2026.02 | 48.2 | — | — | 48.2 | — | — | 36.5 | 15.3 | |
| AFlowOpponent=EmoDx, Evaluator=Human Evaluation, Metric Dimension=Identification2026.02 | 48 | — | 48 | — | — | — | 37.5 | 14.5 | |
| AFlowOpponent=Claude-3.5, Evaluator=GPT-5.2, Metric Dimension=Comfort2026.02 | 46.5 | — | — | 46.5 | — | — | 38 | 15.5 | |
| AFlowOpponent=Claude-3.5, Evaluator=Human Evaluation, Metric Dimension=Overall2026.02 | 46.5 | — | — | — | — | 46.5 | 38 | 15.5 | |
| AFlowOpponent=GPT-4o, Evaluator=Human Evaluation, Metric Dimension=Comfort2026.02 | 45.8 | — | — | 45.8 | — | — | 38.5 | 15.7 | |
| AFlowOpponent=EmoDx, Evaluator=GPT-5.2, Metric Dimension=Comfort2026.02 | 45.5 | — | — | 45.5 | — | — | 38.2 | 16.3 | |
| AFlowOpponent=GPT-4o, Evaluator=Human Evaluation, Metric Dimension=Identification2026.02 | 45.5 | — | 45.5 | — | — | — | 41.5 | 13 | |
| AFlowOpponent=Claude-3.5, Evaluator=Human Evaluation, Metric Dimension=Comfort2026.02 | 44.2 | — | — | 44.2 | — | — | 40 | 15.8 | |
| AFlowOpponent=EmoDx, Evaluator=Human Evaluation, Metric Dimension=Comfort2026.02 | 42.5 | — | — | 42.5 | — | — | 40.5 | 17 | |
| AFlowOpponent=Claude-3.5, Evaluator=GPT-5.2, Metric Dimension=Identification2026.02 | 40.2 | — | 40.2 | — | — | — | 48.5 | 11.3 | |
| AFlowOpponent=Claude-3.5, Evaluator=Human Evaluation, Metric Dimension=Identification2026.02 | 38.5 | — | 38.5 | — | — | — | 50.5 | 11 | |
| AFlowOpponent=ESCoT, Evaluator=GPT-5.2, Metric Dimension=Fluency2026.02 | 35.5 | 35.5 | — | — | — | — | 55.2 | 9.3 | |
| AFlowOpponent=ESCoT, Evaluator=Human Evaluation, Metric Dimension=Fluency2026.02 | 32 | 32 | — | — | — | — | 58.5 | 9.5 | |
| AFlowOpponent=EmoDx, Evaluator=GPT-5.2, Metric Dimension=Fluency2026.02 | 30.5 | 30.5 | — | — | — | — | 62 | 7.5 | |
| AFlowOpponent=EmoDx, Evaluator=Human Evaluation, Metric Dimension=Fluency2026.02 | 28.5 | 28.5 | — | — | — | — | 64 | 7.5 | |
| AFlowOpponent=GPT-4o, Evaluator=GPT-5.2, Metric Dimension=Fluency2026.02 | 15.5 | 15.5 | — | — | — | — | 70.2 | 14.3 | |
| AFlowOpponent=Claude-3.5, Evaluator=GPT-5.2, Metric Dimension=Fluency2026.02 | 12.8 | 12.8 | — | — | — | — | 75.5 | 11.7 | |
| AFlowOpponent=GPT-4o, Evaluator=Human Evaluation, Metric Dimension=Fluency2026.02 | 12.5 | 12.5 | — | — | — | — | 74 | 13.5 | |
| AFlowOpponent=Claude-3.5, Evaluator=Human Evaluation, Metric Dimension=Fluency2026.02 | 10.5 | 10.5 | — | — | — | — | 78 | 11.5 |