Evaluation set
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Evaluation set
97.4ASR
48
Evaluation set (test)
58.5Attack Success Rate (ASR)
48
Evaluation set
93.61Accuracy
15
5-agent 8x8 evaluation set
18Valid Rate
7
Evaluation set without audio prompt (test)
3.98Musicality
7
Evaluation Set
1.844Loss
4