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

Tool Learning under Instructions Beyond Tool Capabilities on NoisyToolBench IBTC 1.0 (test)

98A1 Score

CoT + AwN

-3.9222.544975.46Aug 31, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.08
98
2024.08
98
2024.08
94
2024.08
94
2024.08
94
2024.08
92
2024.08
88
2024.08
86
2024.08
84
2024.08
76
2024.08
70
2024.08
68
2024.08
64
2024.08
62
2024.08
54
2024.08
52
2024.08
48
2024.08
48
2024.08
48
2024.08
48
2024.08
46
2024.08
40
2024.08
38
2024.08
34
2024.08
30
2024.08
30
2024.08
10
2024.08
8
2024.08
0
2024.08
0
2024.08
0
2024.08
0