Temporal Reasoning Prediction on WIKI (test)
99.28Positive PerformanceGETER
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GETERBackbone=Mistral-7B-Instruct2025.05 | 99.28 | 94.49 | 96.19 | 96.81 | |
| GETERBackbone=Llama3-8B-Instruct2025.05 | 98.99 | 90.58 | 91 | 93.79 | |
| GETERBackbone=Qwen2.5-7B-Instruct2025.05 | 97.32 | 93.33 | 94.01 | 95.02 | |
| LoRABackbone=Mistral-7B-Instruct2025.05 | 89.29 | 86.61 | 87.04 | 87.73 | |
| LoRABackbone=Qwen2.5-7B-Instruct2025.05 | 88.65 | 78.58 | 87.36 | 85.19 | |
| LoRABackbone=Llama3-8B-Instruct2025.05 | 88.59 | 73.29 | 81.36 | 81.57 | |
| LoRABackbone=Qwen2.5-7B-Instruct, Reasoning Chain=w/o chains text2025.05 | 87.16 | 80.29 | 87 | 85.04 | |
| LoRABackbone=Mistral-7B-Instruct, Reasoning Chain=w/o chains text2025.05 | 84.94 | 77.82 | 83.08 | 82.18 | |
| LoRABackbone=Llama3-8B-Instruct, Reasoning Chain=w/o chains text2025.05 | 84.08 | 70.67 | 83.36 | 79.8 | |
| GPT-4oType=zero-shot w/o chains text2025.05 | 69.5 | 53.45 | 17.77 | 47.45 | |
| Mistral-7B-InstructType=zero-shot2025.05 | 62.98 | 44.44 | 41.89 | 50.37 | |
| GPT-4oType=zero-shot2025.05 | 61.94 | 37.44 | 40.88 | 47.54 | |
| Llama3-8B-InstructType=zero-shot2025.05 | 51.76 | 26.43 | 1.26 | 27.31 | |
| Llama3-8B-InstructType=zero-shot w/o chains text2025.05 | 48.31 | 54.39 | 66.46 | 52.44 | |
| Qwen2.5-7B-InstructType=zero-shot2025.05 | 46.46 | 47.84 | 2.47 | 32.23 | |
| Qwen2.5-7B-InstructType=zero-shot w/o chains text2025.05 | 43.51 | 53.31 | 7.72 | 34.54 | |
| Mistral-7B-InstructType=zero-shot w/o chains text2025.05 | 35.81 | 49.71 | 55.4 | 46.52 |