Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Model Editing on UltraEditBench

85.7Efficacy

UltraEdit

-3.42819.71142.8565.989May 20, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.05
85.781.2868.73
2025.05
85.781.2868.73
2025.05
85.6981.8865.64
2025.05
84.0376.6264.03
2025.05
84.0376.6264.03
2025.05
83.7177.367.26
2025.05
83.7177.367.26
2025.05
83.4579.1178.05
2025.05
83.4579.1178.05
2025.05
81.777.2577.09
2025.05
81.777.2577.09
2025.05
81.6576.876.44
2025.05
81.6576.876.44
2025.05
81.4275.3562.13
2025.05
80.775.7876.01
2025.05
80.775.7876.01
2025.05
79.0171.4564.1
2025.05
79.0171.4564.1
2025.05
76.570.6861.29
2025.05
52.5147.8847.5
2025.05
47.2144.5539.13
2025.05
47.0838.7639.27
2025.05
42.2741.6539.79
2025.05
40.6434.1837.29
2025.05
22.1912.096.88
2025.05
22.0317.6119.6
2025.05
17.448.015.42
2025.05
13.511.9210.18
2025.05
13.210.5310.27
2025.05
4.513.162.78
2025.05
4.363.484.38
2025.05
3.483.453.43
2025.05
2.422.483.47
2025.05
1.861.551.84
2025.05
1.711.711.83
2025.05
1.210.850.32
2025.05
0.820.320.69
2025.05
0.760.480.78
2025.05
0.740.450.21
2025.05
0.550.090
2025.05
0.320.270.28
2025.05
0.270.060.14
2025.05
0.250.180.2
2025.05
0.090.060.08
2025.05
0.060.360.07
2025.05
000
2025.05
000
2025.05
000
2025.05
000
2025.05
000
2025.05
000