Viewpoint classification on Immigration-3K
0.9174F1 ScoreGPT-4o mini
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GPT-4o miniContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.9174 | 0.9126 | 0.9223 | |
| GPT-4o miniContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.9033 | 0.9108 | 0.8962 | |
| GPT-4o miniContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.903 | 0.9313 | 0.8793 | |
| Mistral Nemo 12BContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.8521 | 0.861 | 0.8438 | |
| Gemma2 9BContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.8478 | 0.9112 | 0.8049 | |
| Gemma2 9BContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.8455 | 0.8855 | 0.8151 | |
| Gemma2 27BContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.8413 | 0.8954 | 0.8032 | |
| Mistral Nemo 12BContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.8384 | 0.8622 | 0.8185 | |
| Gemma2 27BContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.8281 | 0.8568 | 0.805 | |
| Gemma2 9BContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.8279 | 0.8965 | 0.7838 | |
| Gemma2 27BContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.8228 | 0.9096 | 0.772 | |
| Llama 3.1 8BContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.8165 | 0.8337 | 0.8016 | |
| Mistral Nemo 12BContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.8146 | 0.845 | 0.7906 | |
| Llama 3.1 8BContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.7862 | 0.8787 | 0.7365 | |
| Phi 3.5 miniContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.7603 | 0.8666 | 0.7095 | |
| Llama 3.1 8BContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.7331 | 0.7176 | 0.7527 | |
| Phi 3.5 miniContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.6783 | 0.7072 | 0.6589 | |
| Llama 3.2 3BContext setting=Text, Evaluation protocol=Fine-tuning2025.12 | 0.6297 | 0.6925 | 0.6048 | |
| Phi 3.5 miniContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.5791 | 0.8597 | 0.5591 | |
| Llama 3.2 3BContext setting=Text+KG, Evaluation protocol=Fine-tuning2025.12 | 0.5354 | 0.6354 | 0.5321 | |
| Llama 3.2 3BContext setting=KG, Evaluation protocol=Fine-tuning2025.12 | 0.4838 | 0.6953 | 0.5059 |