Math Word Problem Solving on SVAMP hard (test)
48.2AccuracyMSAT-DEDUCTREASONER
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MSAT-DEDUCTREASONERInput Configuration=digit tokenization, Pre-training method=MSAT2023.06 | 48.2 | 3.2 | |
| DEDUCTREASONERInput Configuration=symbolic masks2023.06 | 45 | — | |
| DEDUCTREASONERInput Configuration=digit tokenization2023.06 | 42.8 | -2.2 | |
| MSAT-ROBERTAGENInput Configuration=digit tokenization, Pre-training method=MSAT2023.06 | 36.2 | 5.9 | |
| ROBERTAGENInput Configuration=symbolic masks2023.06 | 30.3 | — | |
| ROBERTAGENInput Configuration=digit tokenization2023.06 | 19.6 | -10.7 |