Question Answering on MGSM benchmark leaderboard · SOTA2 Research

Question Answering on MGSM

90.79AUCSATMD + MSP

Evaluation Results

Method	Links
SATMD + MSP2026.04		90.79	35.09	69.12
SIVR (Ours)2026.04		89.04	17.54	58.13
SIVR2026.04		89.04	17.54	58.13
SATMD + MSP2026.04		86.38	53.97	74.94
SIVR2026.04		84.7	58.78	73.12
SIVR2026.04		84.7	58.78	73.12
Energy2026.04		83.73	58.82	54.66
SAPLMA2026.04		81.03	41.38	48.44
SAPLMA2026.04		81.03	41.38	48.44
Lookback Lens2026.04		79.56	64.56	71.44
TAD2026.04		79.11	62.96	63.11
Lookback Lens2026.04		78.06	66.67	43.78
TAD2026.04		77.37	56.03	55.88
SAPLMA2026.04		76.48	72.53	70.33
SAPLMA2026.04		76.48	72.53	70.33
TAD2026.04		76.28	41.51	51.31
SIVR2026.04		75.65	74.42	69.23
SIVR2026.04		75.65	74.42	69.23
SE2026.04		73.94	63.08	60.17
SATMD + MSP2026.04		72.27	77.89	55.4
Lookback Lens2026.04		71.05	83.33	62.66
SAR2026.04		70.44	69.84	53.94
SAPLMA2026.04		70.34	88.89	57.77
SAPLMA2026.04		70.34	88.89	57.77
Entropy2026.04		69.72	78.14	45.82
Energy2026.04		69.46	80.65	43.5
Perplexity2026.04		68.97	80.35	47.65
Max Prob2026.04		68.87	82.27	47.84
Temp-Scaled2026.04		67.67	83.46	47.03
Energy2026.04		67.14	67.85	55.47
SE2026.04		66.34	88.89	50.44
CoE-R2026.04		66.09	95.85	32.2
SE2026.04		66	76.67	56.88
Entropy2026.04		64.65	93.57	55.57
SAR2026.04		63.71	66.67	52.09
Entropy2026.04		61.45	96.19	38.87
Perplexity2026.04		61.42	94.9	52.78
CoE-C2026.04		60.88	90.1	38.13
Max Prob2026.04		60.49	95.34	51.94
Perplexity2026.04		58.82	96.19	35.32
Max Prob2026.04		58.54	96.19	35.27
Temp-Scaled2026.04		57.33	94.46	48.79
Temp-Scaled2026.04		57.2	96.54	31.35
P(True)2026.04		56.43	95.85	21.02
CoE-R2026.04		54.84	92.9	45.15
CoE-R2026.04		53.07	96.01	34.22
P(True)2026.04		52.28	96.16	31.9
CoE-C2026.04		52.18	94.46	30.61
P(True)2026.04		50.79	96.67	42.17
CoE-C2026.04		46.24	90.24	37.03
SAR2026.04		39.67	97.62	33.7