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

Hallucination Detection on NQ

0.889AUC

Max Pooling

Updated 1mo ago

Evaluation Results

Method
Max Pooling 2026.05	0.889	-	-	-	-
HaMI (SP) 2026.05	0.885	-	-	-	-
Mean Pooling 2026.05	0.87	-	-	-	-
PR 2026.01	0.8645	-	-	-	-
HaMI (SP) 2026.05	0.864	-	-	-	-
HaMI (Original) 2026.05	0.862	-	-	-	-
Max Pooling 2026.05	0.859	-	-	-	-
Mean Pooling 2026.05	0.841	-	-	-	-
HaMI (SP) 2026.05	0.84	-	-	-	-
HaMI (Original) 2026.05	0.83	-	-	-	-
QAODQ⊕V⊥ 2026.05	0.8276	-	-	-	81.76
QAODQ⊕V⊥ 2026.05	0.8269	-	-	-	81.25
PR 2026.01	0.8191	-	-	-	-
Attention probe, soft target 2026.06	0.8152	-	-	-	-
QAODQ⊕V⊥ 2026.05	0.8142	-	-	-	85.68
Max Pooling 2026.05	0.814	-	-	-	-
PR 2026.01	0.8092	-	-	-	-
Pfalse 2026.05	0.803	-	-	-	81.84
QAODV⊥ 2026.05	0.7998	-	-	-	80.25
QAODV⊥ 2026.05	0.7927	-	-	-	80.48
Mean Pooling 2026.05	0.792	-	-	-	-
MoP 2026.01	0.7919	-	-	-	-
PR 2026.01	0.791	-	-	-	-
PR 2026.01	0.791	-	-	-	-
PR 2026.01	0.7909	-	-	-	-
PR 2026.01	0.7909	-	-	-	-
Pfalse 2026.05	0.7905	-	-	-	77.71
Attention probe 2026.06	0.7897	-	-	-	-
SAPLMA 2026.05	0.7893	-	-	-	79.03
SAPLMA 2026.05	0.7878	-	-	-	80.27
SAPLMA 2026.05	0.784	-	-	-	84.96
SINdex 2026.05	0.783	-	-	-	74.65
PR 2026.01	0.7824	-	-	-	-
MoP 2026.01	0.7814	-	-	-	-
QAODQ⊕V⊥ 2026.05	0.7794	-	-	-	81.68
QAODV⊥ 2026.05	0.7791	-	-	-	85.19
HaMI (Original) 2026.05	0.777	-	-	-	-
SINdex 2026.05	0.7733	-	-	-	79.43
Attention probe, soft target 2026.06	0.7728	-	-	-	-
MoP 2026.01	0.7687	-	-	-	-
MoP 2026.01	0.7687	-	-	-	-
Attention probe 2026.06	0.7661	-	-	-	-
MoP 2026.01	0.7638	-	-	-	-
MoP 2026.01	0.7612	-	-	-	-
MoP-VanillaExperts 2026.01	0.7573	-	-	-	-
MoP-VanillaExperts 2026.01	0.753	-	-	-	-
LSC 2026.01	0.752	38.7	-	-	-
SINdex 2026.05	0.752	-	-	-	85.05
QAODV⊥ 2026.05	0.7508	-	-	-	81.2
Attention probe 2026.06	0.75	-	-	-	-
EigenScore 2026.01	0.749	30.4	-	-	-
Probing Baseline 2026.01	0.7483	-	-	-	-
MoP-VanillaExperts 2026.01	0.7478	-	-	-	-
MoP 2026.01	0.7464	-	-	-	-
MoP 2026.01	0.7464	-	-	-	-
Attention probe, soft target 2026.06	0.7443	-	-	-	-
Probing Baseline 2026.01	0.7439	-	-	-	-
Attention probe, soft target 2026.06	0.7436	-	-	-	-
Probing Baseline 2026.01	0.742	-	-	-	-
Attention probe, soft target 2026.06	0.7413	-	-	-	-
Probing Baseline 2026.01	0.741	-	-	-	-
SAPLMA 2026.05	0.74	-	-	-	80.9
MoP-VanillaExperts 2026.01	0.7385	-	-	-	-
SINdex 2026.05	0.7378	-	-	-	79.72
MoP-VanillaExperts 2026.01	0.7377	-	-	-	-
MoP-VanillaExperts 2026.01	0.7377	-	-	-	-
MIND 2026.05	0.7369	-	-	-	78.34
Probing Baseline 2026.01	0.736	-	-	-	-
Probing Baseline 2026.01	0.736	-	-	-	-
LSC 2026.01	0.732	34	-	-	-
Perplexity 2026.01	0.73	19.5	-	-	-
Attention probe 2026.06	0.7294	-	-	-	-
MIND 2026.05	0.7269	-	-	-	78.3
Attention probe 2026.06	0.7264	-	-	-	-
Attention probe, soft target 2026.06	0.7245	-	-	-	-
MIND 2026.05	0.7241	-	-	-	84
Attention probe 2026.06	0.7222	-	-	-	-
Perplexity 2026.01	0.72	24.5	-	-	-
Pik 2026.05	0.7158	-	-	-	83.72
Attention probe, soft target 2026.06	0.7156	-	-	-	-
Linear probe 2026.06	0.7137	-	-	-	-
MoP-VanillaExperts 2026.01	0.7121	-	-	-	-
MoP-VanillaExperts 2026.01	0.7121	-	-	-	-
Attention probe, soft target 2026.06	0.7118	-	-	-	-
LN-Entropy 2026.01	0.703	9.1	-	-	-
Attention probe 2026.06	0.7022	-	-	-	-
Probing Baseline 2026.01	0.702	-	-	-	-
Probing Baseline 2026.01	0.702	-	-	-	-
LN-Entropy 2026.01	0.702	25.4	-	-	-
Linear probe 2026.06	0.7006	-	-	-	-
Pik 2026.05	0.6999	-	-	-	77.61
Pfalse 2026.05	0.6989	-	-	-	83.73
Linear probe 2026.06	0.6968	-	-	-	-
Logit Entropy 2026.05	0.6914	-	-	-	84.44
Attention probe 2026.06	0.6898	-	-	-	-
MoP-RandomGate 2026.01	0.6876	-	-	-	-
Lexical Similarity 2026.01	0.682	14.8	-	-	-
EigenScore 2026.01	0.681	31.9	-	-	-
Pfalse 2026.05	0.6808	-	-	-	79.57
MoP-RandomGate 2026.01	0.6805	-	-	-	-

Showing 100 of 227 rows