Prediction Consistency Evaluation on Streaming Scam Detection Dataset
91.5ConsistencyORACLE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ORACLE2026.05 | 91.5 | 8.5 | 1.42 | |
| Assessor-only2026.05 | 91.2 | 8.8 | 1.46 | |
| BaselineModel API=GPT-5.12026.05 | 69.6 | 30.4 | 5.08 |
Loading the SOTA2 catalog…