Hallucination Evaluation on MMHE
66.6REGHarmonicEval
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| HarmonicEvalBackbone=GPT-4o2024.12 | 66.6 | 76.4 | 73.4 | 77 | 73.4 | |
| FLEURBackbone=LLaVA-1.5-13B2024.12 | 62.9 | 76.4 | 60.9 | 73.9 | 68.5 | |
| CLIP-S2024.12 | 60.1 | 39.7 | 60.9 | 52 | 53.2 | |
| G-VEval2024.12 | 60.1 | 75 | 71.9 | 68.7 | 68.9 | |
| GPT-FLEURBackbone=GPT-4o2024.12 | 60.1 | 75 | 76.5 | 76 | 71.9 | |
| BART-S2024.12 | 56.4 | 20.5 | 60.9 | 57.8 | 48.9 | |
| ROUGE2024.12 | 49 | 30.8 | 56 | 47.9 | 45.9 | |
| BERT-S2024.12 | 46.2 | 33.8 | 62.1 | 53.1 | 48.8 | |
| BLEU2024.12 | 45.3 | 29.4 | 57.3 | 46.8 | 44.7 | |
| METEOR2024.12 | 44.4 | 29.4 | 59.7 | 53.6 | 46.8 | |
| CIDEr2024.12 | 42.5 | 25 | 62.1 | 42.7 | 43.1 |