Semantic Similarity on WIKI (test)
55.52BLEU-4Llama3-8B-Instruct
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Llama3-8B-InstructType=GETER2025.05 | 55.52 | 68.06 | 69.16 | 88.77 | |
| Qwen2.5-7B-InstructType=GETER2025.05 | 55 | 67.49 | 70 | 88.99 | |
| Mistral-7B-InstructType=GETER2025.05 | 54.96 | 67.74 | 69.17 | 88.92 | |
| Qwen2.5-7B-InstructType=LoRA2025.05 | 53.57 | 67.19 | 67.23 | 88.59 | |
| Mistral-7B-InstructType=LoRA2025.05 | 52.61 | 65.4 | 66.8 | 87.97 | |
| Llama3-8B-InstructType=LoRA2025.05 | 52.09 | 65.27 | 66.67 | 87.99 | |
| Qwen2.5-7B-InstructType=LoRA w/o chains text2025.05 | 51.84 | 65.09 | 65.59 | 87.79 | |
| Mistral-7B-InstructType=LoRA w/o chains text2025.05 | 51.58 | 66.32 | 65.29 | 87.96 | |
| Llama3-8B-InstructType=LoRA w/o chains text2025.05 | 48.99 | 63.53 | 63.08 | 87.13 | |
| GPT-4oType=zero-shot2025.05 | 25.98 | 41.77 | 45.52 | 78.69 | |
| Mistral-7B-InstructType=zero-shot2025.05 | 21.25 | 40.05 | 41.43 | 77.27 | |
| Qwen2.5-7B-InstructType=zero-shot2025.05 | 20.21 | 36.19 | 41.52 | 77.61 | |
| Llama3-8B-InstructType=zero-shot2025.05 | 14.67 | 36.67 | 33.43 | 75.85 | |
| GPT-4oType=zero-shot w/o chains text2025.05 | 13.25 | 28.18 | 36.65 | 69.1 | |
| Mistral-7B-InstructType=zero-shot w/o chains text2025.05 | 11.41 | 26.23 | 31.64 | 67.79 | |
| Llama3-8B-InstructType=zero-shot w/o chains text2025.05 | 9.39 | 25.41 | 27.95 | 66.88 | |
| Qwen2.5-7B-InstructType=zero-shot w/o chains text2025.05 | 7.33 | 21.46 | 34.56 | 66.38 |