Complaint Scene Reasoning on ComplaintScene-Text
83.26Avg Scoresft
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| sftBackbone=Qwen2.5-3B2026.05 | 83.26 | 80.55 | 76.92 | 92.31 | |
| preBackbone=Qwen2.5-3B2026.05 | 81.43 | 78.69 | 74.81 | 90.79 | |
| sftBackbone=Qwen2.5-1.5B2026.05 | 79.14 | 75.88 | 72.36 | 89.18 | |
| preBackbone=Qwen2.5-1.5B2026.05 | 76.58 | 73.24 | 69.11 | 87.39 | |
| llmBackbone=Qwen2.5-3B2026.05 | 74.62 | 71.95 | 66.84 | 85.07 | |
| sftBackbone=Qwen2.5-0.5B2026.05 | 71.48 | 68.95 | 65.74 | 79.75 | |
| llmBackbone=Qwen2.5-1.5B2026.05 | 69.83 | 66.41 | 61.27 | 81.81 | |
| preBackbone=Qwen2.5-0.5B2026.05 | 65.31 | 61.02 | 58.64 | 76.27 | |
| llmBackbone=Qwen2.5-0.5B2026.05 | 57.42 | 54.18 | 49.87 | 68.21 |