List your company

Confidence Estimation on VLCB Pooled Aggregate (test)

4.29ECEInternalInspector
2.669613.607324.54535.4827May 11, 2026

Evaluation results

MethodLinks
InternalInspectorLVLM Backbone=Gemma-3-27B-IT2026.054.2919.6883.1774.15
SAPLMALVLM Backbone=Gemma-3-27B-IT2026.054.6119.4283.8375.45
InternalInspectorLVLM Backbone=Qwen3-VL-8B-Instruct2026.055.4316.9589.7579.6
BICRLVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.055.6518.1687.7478.94
BICRLVLM Backbone=DeepSeek-VL22026.056.0217.986.1981.11
BICRLVLM Backbone=Gemma-3-27B-IT2026.056.9819.5685.176.56
BICRLVLM Backbone=Cross-LVLM Average2026.057.0918.4187.4778.63
InternalInspectorLVLM Backbone=DeepSeek-VL22026.057.3918.9584.5479.31
P(I Know)LVLM Backbone=Qwen3-VL-8B-Instruct2026.057.5417.9988.4377.19
CCPSLVLM Backbone=DeepSeek-VL22026.057.722.1476.9770.94
BICRLVLM Backbone=InternVL3.5-14B2026.057.919.0488.0476.39
InternalInspectorLVLM Backbone=Cross-LVLM Average2026.058.319.8884.3174.85
P(I Know)LVLM Backbone=DeepSeek-VL22026.058.519.2684.5878.49
P(I Know)LVLM Backbone=Gemma-3-27B-IT2026.058.8819.8685.0776.08
BICRLVLM Backbone=Qwen3-VL-8B-Instruct2026.058.9117.490.2680.14
CCPSLVLM Backbone=Gemma-3-27B-IT2026.058.9822.1473.3768.49
P(I Know)LVLM Backbone=Cross-LVLM Average2026.059.3119.4386.376.57
SAPLMALVLM Backbone=Qwen3-VL-8B-Instruct2026.0510.5719.4386.4174.22
InternalInspectorLVLM Backbone=InternVL3.5-14B2026.0510.5921.4882.6469.31
P(I Know)LVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0510.8119.8787.0677.32
P(I Know)LVLM Backbone=InternVL3.5-14B2026.0510.8320.1686.3573.78
SAPLMALVLM Backbone=Cross-LVLM Average2026.0512.2221.3281.6372.94
Prompt EnsemblesLVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0512.7723.4977.3168.47
SAPLMALVLM Backbone=DeepSeek-VL22026.0512.8321.5179.6977.14
InternalInspectorLVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0513.7922.3181.4271.87
CCPSLVLM Backbone=InternVL3.5-14B2026.0514.7223.8371.6458.22
CCPSLVLM Backbone=Cross-LVLM Average2026.0515.2727.1572.8363.1
CCPSLVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0516.2721.987672.89
Prompt EnsemblesLVLM Backbone=DeepSeek-VL22026.0516.324.7272.4973.46
SAPLMALVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0516.5222.9982.0172.47
SAPLMALVLM Backbone=InternVL3.5-14B2026.0516.5823.2776.2165.42
Prompt EnsemblesLVLM Backbone=InternVL3.5-14B2026.0516.725.960.2643.37
Prompt EnsemblesLVLM Backbone=Cross-LVLM Average2026.0518.4626.0769.2860
Prompt EnsemblesLVLM Backbone=Qwen3-VL-8B-Instruct2026.0519.825.9666.252.9
Self-ProbingLVLM Backbone=InternVL3.5-14B2026.0521.5124.7176.7770.75
Self-ProbingLVLM Backbone=Qwen3-VL-8B-Instruct2026.0524.3927.477759.49
P(True)LVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0526.2330.9367.9154.73
Prompt EnsemblesLVLM Backbone=Gemma-3-27B-IT2026.0526.7230.2970.1361.82
Self-ProbingLVLM Backbone=Cross-LVLM Average2026.0527.5129.7977.8165.63
Self-ProbingLVLM Backbone=Gemma-3-27B-IT2026.0527.6729.4679.268.39
CCPSLVLM Backbone=Qwen3-VL-8B-Instruct2026.0528.6745.6466.1844.93
Self-ProbingLVLM Backbone=LLaVA-NeXT-Vicuna-13B2026.0528.9229.9781.8767.3
P(True)LVLM Backbone=DeepSeek-VL22026.0534.0737.4368.0952.66
Self-ProbingLVLM Backbone=DeepSeek-VL22026.0535.0737.3574.1962.22
P(True)LVLM Backbone=Cross-LVLM Average2026.0538.0439.872.8455.66
P(True)LVLM Backbone=InternVL3.5-14B2026.0541.1941.4578.0259.49
P(True)LVLM Backbone=Qwen3-VL-8B-Instruct2026.0543.8844.1676.1854.62
P(True)LVLM Backbone=Gemma-3-27B-IT2026.0544.845.0574.0156.81