Hallucination Classification on HaloQuest
60Balanced AccuracyLogistic Regression
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Logistic Regressionfeatures=5 structural descriptors2026.05 | 60 | 60.3 | 68 | |
| Random Chance2026.05 | 55 | 54.8 | 55 | |
| Majority Class2026.05 | 50 | 39.8 | 50 |
Loading the SOTA2 catalog…