List your company

Uncertainty Quantification on MAQA

83.5Hamming AUCStructure-Aware Minimum Bayes Risk Decoding
47.93257.16666.475.634Jan 29, 2026

Evaluation results

MethodLinks
Structure-Aware Minimum Bayes Risk DecodingBackbone Model=Gemma-3-4B2026.0183.582.8
Structure-Aware Minimum Bayes Risk DecodingBackbone Model=Gemma-3-12B2026.0181.679.5
SEBackbone Model=Gemma-3-4B2026.0181.578.6
KLEBackbone Model=Gemma-3-4B2026.0179.380.2
Structure-Aware Minimum Bayes Risk DecodingBackbone Model=Qwen3-4B2026.0179.180.7
SARBackbone Model=Gemma-3-12B2026.0178.878.5
KLEBackbone Model=Gemma-3-12B2026.0178.477.1
SEBackbone Model=Gemma-3-12B2026.0178.376.2
Structure-Aware Minimum Bayes Risk DecodingBackbone Model=Qwen3-30B-A3B2026.0177.775.6
COCOABackbone Model=Gemma-3-12B2026.0176.875.7
SARBackbone Model=Gemma-3-4B2026.0176.480.2
SARBackbone Model=Qwen3-30B-A3B2026.0176.474.8
COCOABackbone Model=Qwen3-30B-A3B2026.017673.7
SEBackbone Model=Qwen3-30B-A3B2026.0175.472.9
COCOABackbone Model=Gemma-3-4B2026.0174.777.3
KLEBackbone Model=Qwen3-30B-A3B2026.0174.672.4
SARBackbone Model=Qwen3-4B2026.0173.573.3
COCOABackbone Model=Qwen3-4B2026.0173.372.9
SEBackbone Model=Qwen3-4B2026.0173.174.7
MSPBackbone Model=Qwen3-30B-A3B2026.0172.971.2
MSPBackbone Model=Gemma-3-12B2026.0172.471.5
KLEBackbone Model=Qwen3-4B2026.0171.672.5
MSPBackbone Model=Qwen3-4B2026.0170.769.8
MSPBackbone Model=Gemma-3-4B2026.0170.472.6
P(True)Backbone Model=Gemma-3-12B2026.0167.365.7
P(True)Backbone Model=Qwen3-4B2026.0159.556.9
P(True)Backbone Model=Gemma-3-4B2026.0156.655.6
P(True)Backbone Model=Qwen3-30B-A3B2026.0149.350.6