Sequential Model Editing on CounterFact T = 300 benchmark leaderboard · SOTA2 Research

Sequential Model Editing on CounterFact T = 300

99.3EfficacyMEMIT

Evaluation Results

Method	Links
MEMIT2026.05		99.3	61	21.5
AlphaEdit2026.05		98.9	73.5	13.7
PMET2026.05		98.8	72.4	12.6
EMMET2026.05		98.8	73	13.5
BetaEdit2026.05		98.8	74	13.9
PMET2026.05		98.8	63	21.6
EMMET2026.05		98.5	74	11.7
AlphaEdit2026.05		98.5	74.1	12.8
PMET2026.05		98.5	72.5	13.5
EMMET2026.05		98.5	69.4	20.2
BetaEdit2026.05		98	73.5	13
MEMIT2026.05		97.8	71	13.8
AlphaEdit2026.05		97.3	71.5	19.4
MEMIT2026.05		97	70.8	13.2
BetaEdit2026.05		96.8	70.4	19
SimIE2026.05		96.4	72.1	12.8
SimIE2026.05		96	68	20.7
SimIE2026.05		95	71.3	11.8
AdaEdit2026.05		90	63	18.5
AdaEdit2026.05		89	58	10
AdaEdit2026.05		88.4	59.5	8.1
PRUNE2026.05		88	60	17.5
PRUNE2026.05		87	54	10.6
PRUNE2026.05		85	55.5	8.3
RECT2026.05		84	58	17
RECT2026.05		82	52	9.5
RECT2026.05		80.4	53.8	8.7
RLEdit2026.05		80	36	5.5
RLEdit2026.05		80	38	14.7
RLEdit2026.05		76.5	34.4	7
FT2026.05		48	5.9	8.7
FT2026.05		45	5	4
FT2026.05		43.6	4.1	3.2
Pre-edited2026.05		0.5	0.5	15.2
Pre-edited2026.05		0.4	0.2	22.3
Pre-edited2026.05		0.3	0.2	15