Machine Unlearning on RWKU canonical ten-subject batch benchmark leaderboard · SOTA2 Research

Machine Unlearning on RWKU canonical ten-subject batch

85.4FQ ScoreJENSUN

Evaluation Results

Method	Links
JENSUN2026.06		85.4	6.2	43.4	37.2	12.4	42	36.1	49.3	70.2	523	46.5
JENSUN2026.06		85.4	—	—	—	—	—	36.1	49.3	70.2	523	46.5
GRADDIFF2026.06		81.9	5.3	44.5	39.2	15.3	41.6	36.7	45.2	69.2	514	43
GRADDIFF2026.06		81.9	—	—	—	—	—	36.7	45.2	69.2	514	43
ATWU2026.06		81.4	5	56.2	51.2	15.8	50.9	22.7	58.7	70.3	519	46
ATWU2026.06		81.4	—	—	—	—	—	22.7	58.7	70.3	519	46
WGA2026.06		78	6.1	51.5	45.4	18.7	48.6	26.1	51.9	69.4	518	41
WGA2026.06		78	—	—	—	—	—	26.1	51.9	69.4	518	41
SATIMP2026.06		77.3	6.9	53.3	46.4	19.3	46.5	29.3	48	69.8	518	40
SATIMP2026.06		77.3	—	—	—	—	—	29.3	48	69.8	518	40
SIMNPO2026.06		51.8	21.8	56.1	34.3	40.9	46.2	29.7	22.1	68.8	529	45.5
SIMNPO2026.06		51.8	—	—	—	—	—	29.7	22.1	68.8	529	45.5
NPO2026.06		47.6	23.6	54.1	30.5	44.4	40.1	39.1	8.5	68.1	543	33.5
NPO2026.06		47.6	—	—	—	—	—	39.1	8.5	68.1	543	33.5
DPO2026.06		12.3	49.9	59	9	74.4	54.6	17	0	69.7	523	43
DPO2026.06		12.3	—	—	—	—	—	17	0	69.7	523	43
ORIGINAL2026.06		0	68.3	75.9	7.7	84.8	65.8	0	0	70.1	529	50
ORIGINAL2026.06		0	—	—	—	—	—	0	0	70.1	529	50