Machine-Generated Text Attribution on AIGTBench
91.5Macro F1 (Full)RidgeFT
Evaluation results
| Method | Links | |
|---|---|---|
| RidgeFTBackbone=RoBERTa-base, New Model=Llama-12026.06 | 91.5 | |
| RidgeFTBackbone=DeBERTa-base, New Model=Llama-12026.06 | 91 | |
| RidgeFTBackbone=RoBERTa-base, New Model=Llama-3.12026.06 | 89.9 | |
| RidgeFTBackbone=RoBERTa-base, New Model=Average2026.06 | 89.9 | |
| RidgeFTBackbone=RoBERTa-base, New Model=GPT-4o-mini2026.06 | 89.6 | |
| RidgeFTBackbone=DeBERTa-base, New Model=Average2026.06 | 89.6 | |
| RidgeFTBackbone=DeBERTa-base, New Model=Llama-3.12026.06 | 89.5 | |
| RidgeFTBackbone=RoBERTa-base, New Model=GPT-3.52026.06 | 89.4 | |
| RidgeFTBackbone=DeBERTa-base, New Model=Llama-22026.06 | 89.3 | |
| RidgeFTBackbone=DeBERTa-base, New Model=GPT-3.52026.06 | 89.2 | |
| RidgeFTBackbone=RoBERTa-base, New Model=Llama-22026.06 | 89.2 | |
| RidgeFTBackbone=DeBERTa-base, New Model=GPT-4o-mini2026.06 | 88.9 |