Inbox Sorting on PMR-Synth (test)
0.73T-NDCG@10Qwen32B_SFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen32B_SFTEvaluation Protocol=UrgentSFT Pairs2026.01 | 0.73 | 0.34 | |
| MedGem27BEvaluation Protocol=0-Shot Pairs2026.01 | 0.66 | 0.32 | |
| MedGem27B_SFTEvaluation Protocol=UrgentSFT Pairs2026.01 | 0.66 | 0.35 | |
| Reward-8B_UrgentEvaluation Protocol=UrgentReward Pairs2026.01 | 0.64 | 0.29 | |
| Qwen32BEvaluation Protocol=0-Shot Pairs2026.01 | 0.62 | 0.28 | |
| Qwen8B_SFTEvaluation Protocol=UrgentSFT Pairs2026.01 | 0.62 | 0.3 | |
| GPT-OSS*Evaluation Protocol=0-Shot Pairs, Parameters=120B2026.01 | 0.58 | 0.28 | |
| Qwen32B_SFTEvaluation Protocol=Multi-Class2026.01 | 0.57 | 0.27 | |
| Reward-8B_BaseEvaluation Protocol=UrgentReward Pairs2026.01 | 0.57 | 0.26 | |
| Reward-4B_UrgentEvaluation Protocol=UrgentReward Pairs2026.01 | 0.55 | 0.26 | |
| MedGem27BEvaluation Protocol=Multi-Class2026.01 | 0.54 | 0.24 | |
| Qwen8BEvaluation Protocol=0-Shot Pairs2026.01 | 0.52 | 0.27 | |
| Qwen32BEvaluation Protocol=Multi-Class2026.01 | 0.5 | 0.26 | |
| GPT-OSSEvaluation Protocol=Multi-Class2026.01 | 0.48 | 0.23 | |
| Reward-4B_BaseEvaluation Protocol=UrgentReward Pairs2026.01 | 0.37 | 0.18 | |
| MedGem27B_SFTEvaluation Protocol=Multi-Class2026.01 | 0.27 | 0.18 |