KAMEL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
KAMEL
16.2Macro EM
19
KAMEL multi-token targets (N=4)
73Efficacy
4
KAMEL multi-token targets (N=3)
76Efficacy
4
KAMEL multi-token targets N=2
88Efficacy
4
KAMEL 4 tokens
95Editing Efficacy
4
KAMEL 3 tokens
96.5Editing Efficacy
4
KAMEL 2 tokens
98.5Editing Efficacy
4
KAMEL 1 token
100Editing Efficacy
4