Sequential Model Editing on ZsRE benchmark leaderboard · SOTA2 Research

Sequential Model Editing on ZsRE

99.74EfficacyAlphaEdit

Evaluation Results

Method	Links
AlphaEdit2025.10		99.74	91.58	41.58
EvoEdit2025.10		99.54	91.67	41.97
NSE2024.10		96.87	91.33	28.66
EvoEdit2025.10		95.74	92.13	32.41
EvoEdit2025.10		94.97	89.92	29.26
MEMIT2024.10		94.91	90.22	27.56
AlphaEdit2025.10		94.47	91.13	32.55
AlphaEdit2025.10		94.44	89.87	30.09
GRACE2024.10		94.33	1.59	27.63
SPHERE2025.10		90.01	84.67	45.4
AlphaEdit2025.10		86.64	81.28	28.78
RECT2025.10		86.05	80.54	31.67
NSE2024.10		83.26	75.33	26.14
GRACE2024.10		82.54	0.4	24.78
MEMIT2024.10		79.17	71.44	26.12
GRACE2024.10		74.58	1.03	31.86
FT-L2024.10		72.37	68.91	19.66
SPHERE2025.10		70.04	65.43	27.35
NSE2024.10		62.29	47.13	32.32
ROME2024.10		56.42	54.65	9.86
PRUNE2025.10		47.93	45.5	39.2
ROME2024.10		47.5	43.56	14.27
AlphaEdit2025.10		42.01	39.99	13.87
FT-W2024.10		39.81	32.55	27.76
FT-L2024.10		37.11	33.3	10.36
Pre-edited2024.10		36.99	36.34	31.89
Pre-edited2025.10		35.42	34.17	38.02
Pre-edited2025.10		35.29	34.1	38.44
MEMIT2024.10		34.62	31.28	18.49
MEMIT2025.10		34.62	31.28	18.49
MEMIT2025.10		34.62	31.28	18.49
FT-W2024.10		32.08	31.43	14.72
FT-L2024.10		30.48	30.22	15.49
FT2025.10		30.48	30.22	15.49
Pre-edited2024.10		26.32	25.79	27.42
FT-W2024.10		24.97	22.4	12.73
PRUNE2025.10		24.77	23.87	20.69
Pre-edited2024.10		22.19	31.3	24.15
ROME2025.10		18.93	17.2	7.56
FT2025.10		15.27	14.78	5.06
PRUNE2025.10		10.35	10.08	9.55
FT2025.10		4.97	4.58	4.01
ROME2025.10		2.25	2.03	0.05
ROME2024.10		2.01	1.8	0.69
ROME2025.10		2.01	1.8	0.69
InstructEdit2025.10		1.58	1.36	1.01
MEND2024.10		0.91	1.09	0.53
MEND2025.10		0.91	1.09	0.53
RECT2025.10		0.73	0.75	0.05
MEND2024.10		0.71	0.71	0.52
MEMIT2025.10		0.13	0.12	0.04
RECT2025.10		0.01	0.01	0.04
MEND2024.10		0	0	0
MEMIT2025.10		0	0	0.06
MEMIT2025.10		0	0	1.59