Symbolic Reasoning on Date Understanding (DU)
87.2AccuracySoftCoT
Evaluation Results
| Method | Links | |
|---|---|---|
| SoftCoTN=10, Backbone=Qwen3-8B2025.02 | 87.2 | |
| SoftCoTN=1, Backbone=Qwen3-8B2025.02 | 85.6 | |
| Zero-Shot Assist-CoTN=10, Backbone=Qwen3-8B2025.02 | 84.8 | |
| Zero-Shot CoTN=10, Backbone=Qwen3-8B2025.02 | 84.56 | |
| Zero-Shot Assist-CoTN=1, Backbone=Qwen3-8B2025.02 | 80.56 | |
| Zero-Shot CoTN=1, Backbone=Qwen3-8B2025.02 | 80.32 | |
| SoftCoTBackbone=LLaMA-3.1-8B-Instruct2025.02 | 59.04 | |
| Zero-Shot Assist-CoTBackbone=LLaMA-3.1-8B-Instruct2025.02 | 58.24 | |
| Zero-Shot CoTBackbone=LLaMA-3.1-8B-Instruct2025.02 | 54.4 | |
| Zero-Shot CoT-UnkBackbone=LLaMA-3.1-8B-Instruct2025.02 | 54.16 |