Share your thoughts, 1 month free Claude Pro on usSee more

Uncertainty Estimation on TruthfulQA

63.9AUROC

Closeness Centrality

Updated 1mo ago

Evaluation Results

Method	Links
Closeness Centrality 2026.04		63.9	-	-
Kernel Lang. Ent. 2026.04		63.7	-	-
Internal Confidence 2025.06		63.2	26.8	15.4
SC Score 2026.04		60.3	-	-
SC Based VC 2026.04		60.2	-	-
Predictive Entropy 2025.06		60	13.7	-
Predictive Entropy 2025.06		59.9	21.1	-
Self Certainty 2026.04		58.6	-	-
Min-K Entropy 2025.06		58.1	22.8	-
SC + VC 2026.04		58.1	-	-
Internal Confidence 2025.06		58	16.4	0.5
Total 2026.04		57.9	-	-
Min-K Entropy 2025.06		57.8	21	-
Internal Confidence 2025.06		56.4	13.2	40.7
Predictive Entropy 2025.06		55.7	15.6	-
Min-K Entropy 2025.06		55.1	13.9	-
Mean Token Entropy 2026.04		54.4	-	-
Perplexity 2025.06		54.3	4.8	-
Token Entropy 2026.04		54.2	-	-
Perplexity 2025.06		54.1	3.8	-
Max Sequence Prob. 2026.04		53.4	-	-
SelfCheckGPT 2026.04		53.4	-	-
Attentional Entropy 2025.06		53.3	9.6	-
Aleatoric 2026.04		52.8	-	-
Max Token Prob. 2026.04		52.8	-	-
Perplexity 2026.04		52.7	-	-
Perplexity 2025.06		52.6	6	-
Attentional Entropy 2025.06		52.4	8.3	-
Max(− log p) 2025.06		52.4	4.1	-
SemanticEntropy 2026.04		51.4	-	-
Max(− log p) 2025.06		51	4.7	-
Attentional Entropy 2025.06		50.7	3.2	-
Max(− log p) 2025.06		50.3	-1.3	-
P(YES) (naive avg) 2025.06		49.7	-3.5	8.6
P(YES) (naive avg) 2025.06		49.3	-2	25.5
P(YES) (naive avg) 2025.06		47	1.9	3.4
PTrue 2026.04		45.1	-	-
P(YES) (top right) 2025.06		43.3	-11.9	55.7
P(YES) (top right) 2025.06		42.6	-19	53.4
P(YES) (top right) 2025.06		39.6	-16.2	27.1