Machine Unlearning on TOFU Avg. benchmark leaderboard · SOTA2 Research

Machine Unlearning on TOFU Avg.

14.32Log Forget Quality (-log(FQ))Ref

Evaluation Results

Method	Links
Ref2026.04		14.32	0	0	99.92
Ref2026.04		13.492	0	0	99.96
Ref2026.04		11.953	0	0	99.8
SCE2026.04		6.262	1.522	-1.23	19.52
SatImp2026.04		5.097	2.57	11.95	26.47
SCN2026.04		4.646	1.358	3.812	58.43
WGA2026.04		3.472	1.375	5.637	52.14
GA2026.04		3.169	1.588	4.321	35.1
WGA2026.04		2.944	3.516	10.338	27.15
SCE2026.04		2.914	5.357	16.019	33.18
SCN2026.04		2.914	5.1	13.964	39.36
TNPO2026.04		2.88	4.803	18.05	47.9
SCE2026.04		2.804	7.008	11.698	21.33
GA2026.04		2.081	8.507	15.558	23.44
GA2026.04		1.975	6.198	10.044	18.92
SatImp2026.04		1.87	2.203	4.721	6.35
TNPO2026.04		1.861	5.313	10.153	32.93
SCN2026.04		1.692	5.732	8.408	47.03
TNPO2026.04		1.341	1.63	2.942	44.81
SatImp2026.04		1.252	5.614	7.344	23.62
ETW2026.04		1.176	2.274	1.873	6.57
WGA2026.04		1.001	5.724	5.992	10.21
NPO2026.04		0.796	3.573	1.644	33.63
NPO2026.04		0.456	7.85	3.179	23.1
ETW2026.04		0.264	6.323	1.387	6.4
ETW2026.04		0.236	4.722	1.099	15.42
NPO2026.04		0.182	3.555	0.85	8.28
Retrain2026.04		0	0.518	0	0.97
Retrain2026.04		0	0.972	0	0.8
Retrain2026.04		0	0.907	0	0.87