TQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
TQA
93.3AUROC
98
TQA 16 additional samples (test)
94.1AUROC
98
TQA 8 additional samples
94AUROC (TQA 8 additional samples)
98
TQA 4 additional samples (test)
0.937AUROC
98
TQA (test)
90.2AUROC
90
TQA
0.173Absolute Execution Time Overhead (s)
90
TQA
86.1PRR
90
TQA
92.3Accuracy
80
TQA
76.8Accuracy
60
TQA
0.072Brier Score
28
TQA D_special (test)
28.46r_pri
20
TQA FinQA, HiTab, TAT-QA, TabMWP, WTQ
40.48FinQA Accuracy
20
TQA Poison Attack (test)
75.6Accuracy
18
TQA PIA Attack (test)
76.4Accuracy
18
TQA
83.2Accuracy
18
TQA poison @ Position 10, k=10 (test)
71Robustness Accuracy
15
TQA poison @ Position 1, k=10 (test)
66.4Robustness Accuracy
15
TQA
42.12EM
14
TQA
77.5Accuracy
13
TQA
0.05Relative Execution Time Overhead
12
TQA (test)
66.45EM
11
TQA (test)
44.13r_connect
10
TQA D_special (test)
44.41r_pub Score
10
TQA
77.3F1 Score
9
TQA
86.7Accuracy
8