Fact Retrieval on FinQA (test)
93.31Recall@3APOLLO (Ensemble model)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| APOLLO (Ensemble model)Pre-training=DeBERTa-v32022.12 | 93.31 | 94.98 | |
| APOLLOPre-training=DeBERTa-v32022.12 | 92.37 | 94.49 | |
| APOLLOPre-training=RoBERTa2022.12 | 91.76 | 93.95 | |
| Ant Risk AI (Ensemble model)Pre-training=RoBERTa2022.12 | 90.77 | 94.33 | |
| FinQANetPre-training=DeBERTa-v32022.12 | 90.37 | 93.78 | |
| Ant Risk AIPre-training=RoBERTa2022.12 | 90.16 | 94.12 | |
| FinQANetPre-training=RoBERTa2022.12 | 89.82 | 93.22 |