Relevance Classification on ESCI-ES (test)
68.1ACCBest-SingleCoT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Best-SingleCoTModel Category=Teacher Model, Protocol=SFT+DPO2026.01 | 68.1 | 58.21 | |
| MPCoTModel Category=Teacher Model, Protocol=SFT+DPO2026.01 | 67.32 | 60.31 | |
| LRKDModel Category=Student Model, Architecture=GAT2026.01 | 66.3 | 55.89 | |
| LRKDModel Category=Student Model, Architecture=Poly2026.01 | 65.89 | 55.47 | |
| MKDModel Category=Student Model2026.01 | 65.64 | 54.93 | |
| LRKDModel Category=Student Model, Architecture=MLP2026.01 | 64.99 | 54.61 | |
| CED-KDModel Category=Student Model2026.01 | 64.51 | 53.61 | |
| BERT-multilingual-baseModel Category=Deployable Model Baseline2026.01 | 64.5 | 53.25 | |
| MPCoTModel Category=Teacher Model, Protocol=SFT2026.01 | 61.37 | 56.02 | |
| ProgressiveCoTModel Category=Teacher Model, Protocol=SFT2026.01 | 61.03 | 54.05 | |
| Best-SingleCoTModel Category=Teacher Model, Protocol=SFT2026.01 | 60.62 | 55.28 | |
| Best-SingleCoTModel Category=Teacher Model, Protocol=few-shot2026.01 | 58.95 | 52.88 | |
| MPCoTModel Category=Teacher Model, Protocol=few-shot2026.01 | 57.62 | 49.49 |