Fake news detection on PolitiFact Adversarial (test)
78.44Macro F1SheepDog
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SheepDogModel Group=G2: Fine-tuned adversarial models2026.01 | 78.44 | 11.48 | |
| LLaMA-3.1-8B-InstructModel Group=G3: Zero-shot LLMs2026.01 | 67.74 | 4.92 | |
| RoBERTaModel Group=G1: Fine-tuned LMs2026.01 | 59.82 | 21.51 | |
| Qwen2.5-7B-InstructModel Group=G3: Zero-shot LLMs2026.01 | 53.25 | 5.15 |