Uncertainty Quantification on OKVQA
79.9AUROCVSE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VSECategory=Entropy, Venue=Ours, Model=Qwen32026.06 | 79.9 | — | |
| SECategory=Entropy, Venue=ICLR’23, Model=Qwen32026.06 | 79 | — | |
| SCG-PrCategory=Consistency, Venue=EMNLP’23, Model=Qwen2.52026.06 | 76.7 | — | |
| VSECategory=Entropy, Venue=Ours, Model=Qwen2.52026.06 | 76.7 | — | |
| SECategory=Entropy, Venue=ICLR’23, Model=Qwen2.52026.06 | 75.8 | — | |
| SNNECategory=Entropy, Venue=ACL’25, Model=Qwen2.52026.06 | 75.1 | — | |
| VSECategory=Entropy, Venue=Ours, Model=LLaVA2026.06 | 74.9 | — | |
| KLECategory=Entropy, Venue=NIPS’25, Model=Qwen2.52026.06 | 74.4 | — | |
| C&UCategory=Consistency, Venue=CVPR’24, Model=Qwen2.52026.06 | 74.2 | — | |
| VL-UCategory=Entropy, Venue=arxiv, Model=Qwen2.52026.06 | 73.1 | — | |
| VL-UCategory=Entropy, Venue=arxiv, Model=Qwen32026.06 | 73.1 | — | |
| AvgEntCategory=Logit, Venue=EMNLP’24, Model=LLaVA2026.06 | 72.9 | — | |
| MaxEntCategory=Logit, Venue=EMNLP’24, Model=LLaVA2026.06 | 72.4 | — | |
| KLECategory=Entropy, Venue=NIPS’25, Model=Qwen32026.06 | 72.4 | — | |
| C&UCategory=Consistency, Venue=CVPR’24, Model=LLaVA2026.06 | 71.8 | — | |
| MaxEntCategory=Logit, Venue=EMNLP’24, Model=Qwen32026.06 | 71.5 | — | |
| VSECategory=Entropy, Venue=Ours, Model=Gemma32026.06 | 71.5 | — | |
| MaxEntCategory=Logit, Venue=EMNLP’24, Model=Qwen2.52026.06 | 71.4 | — | |
| SCG-NLICategory=Consistency, Venue=EMNLP’23, Model=Qwen2.52026.06 | 71.4 | — | |
| SNNECategory=Entropy, Venue=ACL’25, Model=Qwen32026.06 | 71.1 | — | |
| AvgProbCategory=Logit, Venue=EMNLP’24, Model=LLaVA2026.06 | 70.5 | — | |
| AvgEntCategory=Logit, Venue=EMNLP’24, Model=Qwen32026.06 | 70.3 | — | |
| SECategory=Entropy, Venue=ICLR’23, Model=Gemma32026.06 | 70.3 | — | |
| VL-UCategory=Entropy, Venue=arxiv, Model=Gemma32026.06 | 70.2 | — | |
| SNNECategory=Entropy, Venue=ACL’25, Model=LLaVA2026.06 | 69.8 | — | |
| VL-UCategory=Entropy, Venue=arxiv, Model=LLaVA2026.06 | 69.5 | — | |
| C&UCategory=Consistency, Venue=CVPR’24, Model=Qwen32026.06 | 69.3 | — | |
| KLEUncertainty Level=Semantic-Level2026.05 | 69.1 | 7.7 | |
| AvgEntCategory=Logit, Venue=EMNLP’24, Model=Qwen2.52026.06 | 69 | — | |
| SECategory=Entropy, Venue=ICLR’23, Model=LLaVA2026.06 | 68.5 | — | |
| AvgProbCategory=Logit, Venue=EMNLP’24, Model=Qwen32026.06 | 67.9 | — | |
| C&UCategory=Consistency, Venue=CVPR’24, Model=Gemma32026.06 | 67.8 | — | |
| SCG-PrCategory=Consistency, Venue=EMNLP’23, Model=LLaVA2026.06 | 67.5 | — | |
| KLECategory=Entropy, Venue=NIPS’25, Model=Gemma32026.06 | 67.3 | — | |
| KLECategory=Entropy, Venue=NIPS’25, Model=LLaVA2026.06 | 67.2 | — | |
| SEUncertainty Level=Semantic-Level2026.05 | 66.9 | 7.8 | |
| SNNECategory=Entropy, Venue=ACL’25, Model=Gemma32026.06 | 66.7 | — | |
| Verb-UCategory=Verbalized, Venue=NAACL’24, Model=Qwen2.52026.06 | 66.3 | — | |
| SCG-NLICategory=Consistency, Venue=EMNLP’23, Model=LLaVA2026.06 | 65.8 | — | |
| SCG-PrCategory=Consistency, Venue=EMNLP’23, Model=Gemma32026.06 | 65.7 | — | |
| SCG-PrCategory=Consistency, Venue=EMNLP’23, Model=Qwen32026.06 | 65.7 | — | |
| VIG-TUQUncertainty Level=Token-Level2026.05 | 65.5 | 10.4 | |
| AvgProbCategory=Logit, Venue=EMNLP’24, Model=Qwen2.52026.06 | 65.3 | — | |
| VIG-TUQ-AUncertainty Level=Token-Level2026.05 | 64.9 | 10.4 | |
| Verb-UCategory=Verbalized, Venue=NAACL’24, Model=Gemma32026.06 | 64.7 | — | |
| VIG-TUQ-JSDUncertainty Level=Token-Level2026.05 | 63.3 | 12.1 | |
| Token-EntropyUncertainty Level=Token-Level2026.05 | 62.9 | 11.4 | |
| CCPUncertainty Level=Token-Level2026.05 | 62.8 | 11.9 | |
| Log-PerplexityUncertainty Level=Token-Level2026.05 | 62.5 | 11.7 | |
| Max-ProbUncertainty Level=Token-Level2026.05 | 62.3 | 12.6 | |
| SCG-NLICategory=Consistency, Venue=EMNLP’23, Model=Qwen32026.06 | 59.7 | — | |
| Verb-UCategory=Verbalized, Venue=NAACL’24, Model=Qwen32026.06 | 59.3 | — | |
| Token-SarUncertainty Level=Token-Level2026.05 | 59.2 | 13.9 | |
| SCG-NLICategory=Consistency, Venue=EMNLP’23, Model=Gemma32026.06 | 58.4 | — | |
| AvgEntCategory=Logit, Venue=EMNLP’24, Model=Gemma32026.06 | 57.7 | — | |
| AvgProbCategory=Logit, Venue=EMNLP’24, Model=Gemma32026.06 | 57 | — | |
| MaxEntCategory=Logit, Venue=EMNLP’24, Model=Gemma32026.06 | 56.7 | — | |
| MaxProbCategory=Logit, Venue=EMNLP’24, Model=Qwen32026.06 | 56.5 | — | |
| MaxProbCategory=Logit, Venue=EMNLP’24, Model=Gemma32026.06 | 55.5 | — | |
| MaxProbCategory=Logit, Venue=EMNLP’24, Model=LLaVA2026.06 | 54.2 | — | |
| Verb-UCategory=Verbalized, Venue=NAACL’24, Model=LLaVA2026.06 | 53.2 | — | |
| MaxProbCategory=Logit, Venue=EMNLP’24, Model=Qwen2.52026.06 | 51.9 | — |