Early Rumor Detection on PHEME (macro-F1, ER)
0.646Macro F1Imitation Agent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Imitation AgentBase LLM=ChatGPT2025.12 | 0.646 | 0.205 | |
| ChatGPTStrategy=Earliest Rumor Detection, Input=First post2025.12 | 0.604 | — | |
| Imitation AgentBase LLM=Mistral2025.12 | 0.59 | 0.276 | |
| MistralStrategy=Earliest Rumor Detection, Input=First post2025.12 | 0.571 | — | |
| HEARD2025.12 | 0.501 | 0.266 | |
| ERD2025.12 | 0.487 | 1 | |
| Imitation AgentBase LLM=Llama32025.12 | 0.478 | 0.283 | |
| Llama3Strategy=Earliest Rumor Detection, Input=First post2025.12 | 0.457 | — | |
| CED2025.12 | 0.453 | 0.978 |