Sequential Knowledge Editing on CounterFact (10,000 Samples) benchmark leaderboard · SOTA2 Research

Sequential Knowledge Editing on CounterFact (10,000 Samples)

99.5Efficacy SuccessREVIVE

Evaluation Results

Method	Links
REVIVE2026.01		99.5	93.92	67.35	600.64	40.63
REVIVE2026.01		98.89	92.28	65.72	618.66	32.74
REVIVE2026.01		98.74	90.08	60.19	615.97	32.66
REVIVE2026.01		97.63	87.76	66.52	616.47	40.69
AlphaEdit2026.01		96.51	86.76	60.8	544.18	19.33
NSE+REVIVE2026.01		96.12	84.49	64.17	592.71	37.74
REVIVE2026.01		95.62	84.6	62.17	603.22	29.39
REVIVE2026.01		94.96	77.27	67.78	612.76	38.69
REVIVE2026.01		94.88	89.49	64.06	608.12	40.18
AlphaEdit+REVIVE2026.01		94.48	78.7	62.87	587.94	38.51
REVIVE2026.01		92.69	79.95	63.09	600.13	29.28
AlphaEdit2026.01		92.13	76.8	56.85	581.49	31.72
MEMIT+REVIVE2026.01		90.46	75.88	63.83	598.21	34.32
NSE2026.01		88.95	69.69	75.46	611.35	33.31
REVIVE2026.01		86.95	81.03	64.21	583.05	35.73
RECT2026.01		86.52	69.5	55.71	499.64	11.41
RECT+REVIVE2026.01		82.99	69.2	65.6	595.69	34.05
PRUNE+REVIVE2026.01		82	70.9	62.82	600.99	34.55
REVIVE2026.01		80.57	69.54	54.76	570.85	28.49
NSE2026.01		77.59	44.42	86.12	607.86	23.31
MEMIT2026.01		70.56	62.42	55.94	516.26	8.74
NSE2026.01		69.22	54.54	69.26	596.41	28.87
RECT2026.01		63.6	55.33	56.69	404.13	4.49
AlphaEdit2026.01		62.48	56.9	52.31	505.5	4.25
MEMIT2026.01		62.3	55.02	48.11	522.1	4.4
RECT2026.01		60.23	54.9	50.56	441.61	5.08
PRUNE2026.01		59.98	55.72	48.56	571.27	1.89
PRUNE2026.01		57.61	54.01	52.87	596.56	6.93
MEMIT2026.01		54.03	52.66	53.63	594.16	5.17
PRUNE2026.01		52.92	51.47	53.91	576.95	5.14
GPT2-XL2026.01		21.82	24.16	78.32	626.69	31.34
GPT-J2026.01		15.22	17.65	83.5	622.01	29.61
LLaMA32026.01		7.02	9.44	89.73	635.47	24.24