Explanation Generation on GDELT (test)
34.46BLEU-4GETER
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GETERbackbone=Llama3-8B-Instruct2025.05 | 34.46 | 55.42 | 51.75 | 83.62 | |
| Llama3-8B-Instructtraining_protocol=LoRA, reasoning_chains=true2025.05 | 34.32 | 54.84 | 51.49 | 83.75 | |
| GETERbackbone=Qwen2.5-7B-Instruct2025.05 | 33.81 | 54.76 | 50.18 | 83.59 | |
| GETERbackbone=Mistral-7B-Instruct2025.05 | 32.18 | 53.27 | 49.06 | 82.83 | |
| Mistral-7B-Instructtraining_protocol=LoRA, reasoning_chains=true2025.05 | 30.93 | 52.24 | 47.28 | 82.28 | |
| Qwen2.5-7B-Instructtraining_protocol=LoRA, reasoning_chains=true2025.05 | 26.1 | 47.3 | 43.85 | 79.93 | |
| Mistral-7B-Instructtraining_protocol=LoRA, reasoning_chains=false2025.05 | 18.22 | 38.08 | 35.74 | 76.76 | |
| Llama3-8B-Instructtraining_protocol=LoRA, reasoning_chains=false2025.05 | 18.12 | 37.05 | 35.92 | 77.51 | |
| Qwen2.5-7B-Instructtraining_protocol=LoRA, reasoning_chains=false2025.05 | 17.15 | 36.89 | 34.52 | 75.71 | |
| GPT-4otraining_protocol=zero-shot, reasoning_chains=true2025.05 | 9.16 | 27.61 | 32.32 | 70.91 | |
| Qwen2.5-7B-Instructtraining_protocol=zero-shot, reasoning_chains=true2025.05 | 7.55 | 26.9 | 25.97 | 70 | |
| Mistral-7B-Instructtraining_protocol=zero-shot, reasoning_chains=true2025.05 | 7.46 | 27.96 | 25.99 | 70.43 | |
| GPT-4otraining_protocol=zero-shot, reasoning_chains=false2025.05 | 5.95 | 21.3 | 26.84 | 64.73 | |
| Llama3-8B-Instructtraining_protocol=zero-shot, reasoning_chains=true2025.05 | 5.61 | 27.1 | 25.73 | 67.42 | |
| Mistral-7B-Instructtraining_protocol=zero-shot, reasoning_chains=false2025.05 | 4.89 | 18.2 | 25.78 | 63.6 | |
| Qwen2.5-7B-Instructtraining_protocol=zero-shot, reasoning_chains=false2025.05 | 3.76 | 17.9 | 28.25 | 63.15 | |
| Llama3-8B-Instructtraining_protocol=zero-shot, reasoning_chains=false2025.05 | 2.38 | 13.41 | 17.03 | 56.98 |