WHEN2TOOL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
WHEN2TOOL single-hop 1.0 (test)
94.3Accuracy
90
WHEN2TOOL Overall
-1Δ Accuracy
7
WHEN2TOOL Hard
-0.8Delta Accuracy (ΔAcc)
7
WHEN2TOOL Medium
-0.7Delta Accuracy
7
WHEN2TOOL Easy
-0.3ΔAcc
7