Hallucination Detection in Image Recontextualization on VIGIL 1.0 (test)
34.14Clothing Macro F1VIGIL
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VIGILevaluation_protocol=LLM-as-a-Judge, parameter_selection=best set of parameters on test category2026.02 | 34.14 | 14.22 | 10.42 | 3.54 | 18.63 | |
| Gemini 2.5 Flashevaluation_protocol=LLM-as-a-Judge, parameter_selection=best set of parameters on test category2026.02 | 22.51 | 17.02 | 6.27 | 12.16 | 19.07 | |
| Qwen3-VL-8B-Instructevaluation_protocol=LLM-as-a-Judge, parameter_selection=best set of parameters on test category2026.02 | 11.36 | 11.03 | 2.37 | 2.55 | 2.24 | |
| Gemma 3 27B ITevaluation_protocol=LLM-as-a-Judge, parameter_selection=best set of parameters on test category2026.02 | 8.61 | 8.19 | 0.25 | 1.46 | 10.4 |