Fact-checking on Check-COVID (120 claims)
85.83AccuracyGPT-5-mini + RAG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5-mini + RAGMode=Single-call, Average evidence pool size=19.32026.03 | 85.83 | 85.71 | |
| PROCLAIMDecision logic=Majority Vote, Average evidence pool size=67.52026.03 | 81.67 | 81.65 | |
| DeepSeek-v3.2 + RAGMode=Single-call, Average evidence pool size=18.52026.03 | 80 | 79.72 | |
| Standard MADAverage evidence pool size=5.02026.03 | 71.67 | 70.68 |