List your company

Uncertainty Estimation on AmbigQA

78.5AUROCKernel Lang. Ent.
56.03661.86867.773.532Apr 18, 2026

Evaluation results

MethodLinks
Kernel Lang. Ent.Backbone=Mistral-7B2026.0478.5
TotalBackbone=Mistral-7B2026.0476.8
SelfCheckGPTBackbone=Mistral-7B2026.0473.3
AleatoricBackbone=Mistral-7B2026.0471.6
Closeness CentralityBackbone=Mistral-7B2026.0468.3
SemanticEntropyBackbone=Mistral-7B2026.0467.8
SC + VCBackbone=Mistral-7B2026.0467.1
PerplexityBackbone=Mistral-7B2026.0466.1
SC Based VCBackbone=Mistral-7B2026.0465.8
Max Token Prob.Backbone=Mistral-7B2026.0465.8
Mean Token EntropyBackbone=Mistral-7B2026.0465.4
Token EntropyBackbone=Mistral-7B2026.0465.4
Max Sequence Prob.Backbone=Mistral-7B2026.0465.1
SC ScoreBackbone=Mistral-7B2026.0464.9
PTrueBackbone=Mistral-7B2026.0460.4
Self CertaintyBackbone=Mistral-7B2026.0456.9