Physical Commonsense Reasoning on PIQA (Accuracy (%), Avg.)

81.08Accuracy (%)

MoNE

Updated 1mo ago

Evaluation Results

Method
MoNE 2026.06	81.08	63.97
He et al. 2026.06	81.08	63.97
Baseline 2026.06	80.52	65.62
Baseline 2026.06	80.2	68.6
MoNE 2026.06	79.76	65.65
Wanda 2026.06	79.65	66.62
OursQ 2026.06	79.6	63.78
EAC-MoE 2026.06	79.54	65.3
OursQ 2026.06	79.54	67.61
Wanda 2026.06	79.38	69.2
PuzzleMoE 2026.06	78.7	62.36
Wanda 2026.06	78.51	61.56
He et al. 2026.06	78.5	59.7
PuzzleMoE 2026.06	78.4	62.08
Attribution-guided and Coverage-Maximized Expert-wise Pruning 2026.06	78.4	64.92
He et al. 2026.06	78.2	53.88
Attribution-guided and Coverage-Maximized Expert-wise Pruning 2026.06	77.97	62.34
Wanda 2026.06	77.86	60.82
C-Prune 2026.06	77.86	61.81
EAC-MoE 2026.06	77.2	62.7
FP 2026.06	75.79	71.01
FP 2026.06	75.41	70.01
MoNE 2026.06	75.3	59.8
FAIR-Calib 2026.06	73.12	66.66
FAIR-Calib 2026.06	73.07	64.64
QuaRot 2026.06	72.52	57.49
FlatQuant 2026.06	71.82	63.98
FlatQuant 2026.06	71.65	62.08
QuaRot 2026.06	71.11	61.14
He et al. 2026.06	68.17	44.17
Wanda 2026.06	64.15	49.29
RTN 2026.06	63.33	45.76
RTN 2026.06	61.7	42
C-Prune 2026.06	58.92	42.39
MoNE 2026.06	53.37	33.94
MoNE 2026.06	49.51	33.24