Personalized Emotional Support on ReddiSupp
98.29Win RateMindTailor
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MindTailorBackbone Model=Gemini, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 98.29 | 1.71 | 0 | |
| MindTailorBackbone Model=Gemini, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 97.93 | 2.07 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=ES-VR, Judge Model=gpt-4o-mini2026.06 | 97.81 | 2.19 | 0 | |
| MindTailorBackbone Model=Gemini, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 97.77 | 2.19 | 4 | |
| MindTailorBackbone Model=Gemma4, Baseline=ES-VR, Judge Model=deepseek-v4-flash2026.06 | 96.12 | 3.86 | 2 | |
| MindTailorBackbone Model=Gemma4, Baseline=ES-VR, Judge Model=glm-4.5-air2026.06 | 95.34 | 4.64 | 2 | |
| MindTailorBackbone Model=Mistral, Baseline=ES-VR, Judge Model=deepseek-v4-flash2026.06 | 94.9 | 5.1 | 0 | |
| MindTailorBackbone Model=Gemma3, Baseline=ES-VR, Judge Model=gpt-4o-mini2026.06 | 92.44 | 7.56 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=ES-VR, Judge Model=glm-4.5-air2026.06 | 91.31 | 8.69 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=ES-VR, Judge Model=glm-4.5-air2026.06 | 90.6 | 9.4 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 90.41 | 9.59 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 89.87 | 10.13 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=ES-VR, Judge Model=gpt-4o-mini2026.06 | 89.79 | 10.21 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 87.93 | 12.05 | 2 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 86.93 | 12.99 | 8 | |
| MindTailorBackbone Model=Gemma4, Baseline=ES-VR, Judge Model=gpt-4o-mini2026.06 | 86.49 | 13.51 | 0 | |
| MindTailorBackbone Model=Gemma4, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 86.45 | 13.53 | 2 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 85.32 | 14.68 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=ES-VR, Judge Model=deepseek-v4-flash2026.06 | 84.57 | 15.43 | 0 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=ES-VR, Judge Model=deepseek-v4-flash2026.06 | 84.54 | 15.43 | 2 | |
| MindTailorBackbone Model=Gemma3, Baseline=ES-VR, Judge Model=deepseek-v4-flash2026.06 | 84.25 | 15.75 | 0 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=ES-VR, Judge Model=glm-4.5-air2026.06 | 84.21 | 15.79 | 0 | |
| MindTailorBackbone Model=Gemma4, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 82.81 | 16.79 | 40 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 79.76 | 20.22 | 2 | |
| MindTailorBackbone Model=Gemma3, Baseline=ES-VR, Judge Model=glm-4.5-air2026.06 | 79.62 | 20.36 | 2 | |
| MindTailorBackbone Model=Mistral, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 79.41 | 20.57 | 2 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=ES-VR, Judge Model=gpt-4o-mini2026.06 | 79.01 | 20.99 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 78.22 | 21.78 | 0 | |
| MindTailorBackbone Model=Gemma3, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 76.75 | 23.25 | 0 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 75.36 | 24.64 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 75.29 | 24.69 | 2 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 75.15 | 24.83 | 2 | |
| MindTailorBackbone Model=Gemini, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 74.06 | 25.65 | 29 | |
| MindTailorBackbone Model=Gemma3, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 73.71 | 26.11 | 19 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 73.48 | 26.5 | 2 | |
| MindTailorBackbone Model=Gemma3, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 72.45 | 27.55 | 0 | |
| MindTailorBackbone Model=Gemma3, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 70.59 | 29.41 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 70.3 | 29.7 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 70.13 | 29.87 | 0 | |
| MindTailorBackbone Model=Gemini, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 66.42 | 33.56 | 2 | |
| MindTailorBackbone Model=Mistral, Baseline=Vanilla, Judge Model=glm-4.5-air2026.06 | 65.73 | 34.27 | 0 | |
| MindTailorBackbone Model=Gemma4, Baseline=Vanilla, Judge Model=gpt-4o-mini2026.06 | 62.43 | 37.57 | 0 | |
| MindTailorBackbone Model=Qwen2.5 72B, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 61.28 | 38.68 | 4 | |
| MindTailorBackbone Model=Gemma3, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 60.3 | 39.58 | 13 | |
| MindTailorBackbone Model=Gemma3, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 59.63 | 40.37 | 0 | |
| MindTailorBackbone Model=Gemma4, Baseline=MentalAgora, Judge Model=glm-4.5-air2026.06 | 57.21 | 42.71 | 8 | |
| MindTailorBackbone Model=Gemma4, Baseline=MentalAgora, Judge Model=deepseek-v4-flash2026.06 | 56.83 | 42.77 | 40 | |
| MindTailorBackbone Model=Qwen2.5 14B, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 55.58 | 44.36 | 6 | |
| MindTailorBackbone Model=Gemma4, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 52.92 | 47.08 | 0 | |
| MindTailorBackbone Model=Mistral, Baseline=Vanilla, Judge Model=deepseek-v4-flash2026.06 | 52.55 | 47.45 | 0 | |
| MindTailorBackbone Model=Gemini, Baseline=MentalAgora, Judge Model=gpt-4o-mini2026.06 | 51.11 | 48.89 | 0 |