List your company

Uncertainty Estimation on HumanEval

0.886AUROCTop-5 + Tests + Pseudo
0.35560.49330.6310.7687Jun 8, 2026

Evaluation results

MethodLinks
Top-5 + Tests + PseudoModel=Devstral-Small-25052026.060.8860.99
Top-5 + Tests + PseudoBackbone=Devstral-Small-25052026.060.8860.99
Generated Tests (FUNC)Model=DeepSeek-Coder-V22026.060.8610.952
Generated Tests (FUNC)Backbone=DeepSeek-Coder-V22026.060.8610.952
Top-5 + Tests + PseudoModel=Qwen3-14B2026.060.8520.983
Top-5 + Tests + PseudoBackbone=Qwen3-14B2026.060.8520.983
Top-5 + Tests + PseudoModel=DeepSeek-Coder-V22026.060.8420.957
Top-5 + Tests + PseudoBackbone=DeepSeek-Coder-V22026.060.8420.957
Generated Tests (FUNC)Model=Qwen3-14B2026.060.8220.958
Generated Tests (FUNC)Backbone=Qwen3-14B2026.060.8220.958
Generated Tests (FUNC)Model=Devstral-Small-25052026.060.8190.94
Generated Tests (FUNC)Backbone=Devstral-Small-25052026.060.8190.94
Pseudo Consistency (ALGO)Model=Devstral-Small-25052026.060.7680.939
Pseudo Consistency (ALGO)Backbone=Devstral-Small-25052026.060.7680.939
Pseudo Consistency (ALGO)Model=DeepSeek-Coder-V22026.060.7590.832
Pseudo Consistency (ALGO)Backbone=DeepSeek-Coder-V22026.060.7590.832
Generated Tests (FUNC)Model=Qwen3-32B2026.060.7560.938
Top-5 Entropy (LEX)Model=Devstral-Small-25052026.060.7490.934
Top-5 Entropy (LEX)Backbone=Devstral-Small-25052026.060.7490.934
Top-5 + Tests + PseudoModel=Qwen3-Coder-30B2026.060.7460.98
Consistency (VR)Model=Qwen3-14B2026.060.7410.955
Consistency (VR)Backbone=Qwen3-14B2026.060.7410.955
Top-5 + Tests + PseudoModel=Qwen3-32B2026.060.7340.944
Generated Tests (FUNC)Model=Qwen3-Coder-30B2026.060.7340.973
Top-5 Entropy (LEX)Model=Qwen3-14B2026.060.7290.952
Top-5 Entropy (LEX)Backbone=Qwen3-14B2026.060.7290.952
Pseudo Consistency (ALGO)Model=Qwen3-14B2026.060.7180.947
Pseudo Consistency (ALGO)Backbone=Qwen3-14B2026.060.7180.947
Mean EntropyModel=Qwen3-14B2026.060.7150.941
Mean EntropyBackbone=Qwen3-14B2026.060.7150.941
Consistency (BLEU)Model=Qwen3-Coder-30B2026.060.7050.981
Consistency (BLEU)Model=Qwen3-14B2026.060.6940.945
Consistency (BLEU)Backbone=Qwen3-14B2026.060.6940.945
Max ProbabilityModel=Devstral-Small-25052026.060.680.902
Max EntropyModel=Qwen3-Coder-30B2026.060.6780.977
Mean ProbabilityModel=Devstral-Small-25052026.060.6750.889
Symb. ClusteringModel=Qwen3-14B2026.060.6730.943
Symb. ClusteringBackbone=Qwen3-14B2026.060.6730.943
Pseudo Consistency (ALGO)Model=Qwen3-Coder-30B2026.060.6710.973
Symb. ClusteringModel=Devstral-Small-25052026.060.6680.893
Symb. ClusteringBackbone=Devstral-Small-25052026.060.6680.893
Max ProbabilityModel=Qwen3-Coder-30B2026.060.6670.96
Mean ProbabilityModel=Qwen3-Coder-30B2026.060.6620.951
Symb. ClusteringModel=Qwen3-32B2026.060.6540.922
Symb. ClusteringModel=DeepSeek-Coder-V22026.060.6460.895
Symb. ClusteringBackbone=DeepSeek-Coder-V22026.060.6460.895
Consistency (VR)Model=Devstral-Small-25052026.060.6440.875
Consistency (VR)Backbone=Devstral-Small-25052026.060.6440.875
Consistency (VR)Model=DeepSeek-Coder-V22026.060.6350.877
Consistency (VR)Backbone=DeepSeek-Coder-V22026.060.6350.877
Max ProbabilityModel=Qwen3-14B2026.060.630.935
Pseudo Consistency (ALGO)Model=Qwen3-32B2026.060.6210.915
Mean ProbabilityModel=Qwen3-14B2026.060.6190.902
Max EntropyModel=Devstral-Small-25052026.060.6190.881
Consistency (BLEU)Model=Devstral-Small-25052026.060.6160.885
Consistency (BLEU)Backbone=Devstral-Small-25052026.060.6160.885
Top-5 Entropy (LEX)Model=Qwen3-32B2026.060.6130.902
Mean EntropyModel=Devstral-Small-25052026.060.6120.876
Mean EntropyBackbone=Devstral-Small-25052026.060.6120.876
Max EntropyModel=Qwen3-14B2026.060.6010.898
Top-5 Entropy (LEX)Model=Qwen3-Coder-30B2026.060.5910.967
Consistency (BLEU)Model=DeepSeek-Coder-V22026.060.5860.85
Consistency (BLEU)Backbone=DeepSeek-Coder-V22026.060.5860.85
Mean EntropyModel=Qwen3-Coder-30B2026.060.5810.957
Consistency (VR)Model=Qwen3-Coder-30B2026.060.5780.962
Consistency (VR)Model=Qwen3-32B2026.060.5740.898
Max ProbabilityModel=Qwen3-32B2026.060.5490.899
Mean ProbabilityModel=Qwen3-32B2026.060.530.9
Max EntropyModel=Qwen3-32B2026.060.5230.889
Symb. ClusteringModel=Qwen3-Coder-30B2026.060.5170.96
Consistency (BLEU)Model=Qwen3-32B2026.060.5160.885
Mean EntropyModel=Qwen3-32B2026.060.490.868
Top-5 Entropy (LEX)Model=DeepSeek-Coder-V22026.060.4660.827
Top-5 Entropy (LEX)Backbone=DeepSeek-Coder-V22026.060.4660.827
Max ProbabilityModel=DeepSeek-Coder-V22026.060.4250.788
Max EntropyModel=DeepSeek-Coder-V22026.060.4120.826
Mean ProbabilityModel=DeepSeek-Coder-V22026.060.3930.767
Mean EntropyModel=DeepSeek-Coder-V22026.060.3760.798
Mean EntropyBackbone=DeepSeek-Coder-V22026.060.3760.798