Robustness Evaluation on Stress Tests
58.1Quantity Stress ScoreSAT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SATTraining Strategy=Structure-Aware Training, Backbone=Qwen-2.5-32B-Instruct2026.01 | 58.1 | 63 | 60.6 | |
| Know. AugTraining Strategy=Knowledge-Based Augmentation, Backbone=Qwen-2.5-32B-Instruct2026.01 | 31 | 35.7 | 33.4 | |
| Ans. AugTraining Strategy=Answer-Based Augmentation, Backbone=Qwen-2.5-32B-Instruct2026.01 | 20.1 | 41.6 | 30.9 | |
| VanillaTraining Strategy=Vanilla, Backbone=Qwen-2.5-32B-Instruct2026.01 | 8.2 | 4.6 | 6.4 |