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

LogiQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Logical ReasoningLogiQA
LogiQA Accuracy78.9
251
Logical ReasoningLogiQA (test)
Accuracy86
151
Logical ReasoningLogiQA-2
Accuracy83.8
116
Logical ReasoningLogiQA
Accuracy80.4
100
Logical ReasoningLogiQA
Accuracy50.23
98
Logical ReasoningLogiQA (val)
Accuracy58.37
50
Logical ReasoningLogiQA (dev)
Accuracy47.3
40
Logical ReasoningLogiQA
Accuracy60.22
34
Logical InferenceLogiQA
Task Success Rate (TSR)76.75
30
Logical ReasoningLogiQA original (test)
Accuracy43.16
22
Confidence alignmentLogiQA
ECE0.039
21
Commonsense ReasoningLogiQA
Accuracy29.8
21
Logical ReasoningLogiQA
Accuracy50.94
20
Logical ReasoningLogiQA
Acc@t146.6
20
Confidence CalibrationLogiQA (out-of-distribution)
ECE8
18
Logical ReasoningLogiQA
Pass@1 Accuracy0.88
18
Correctness PredictionLogiQA
Accuracy67.75
18
Question AnsweringLogiQA
Accuracy44.29
17
Logical ReasoningLogiQA
Pass@1 Accuracy48.61
14
Logical ReasoningLogiQA
Accuracy (LogiQA)68.9
12
Question AnsweringLogiQA (test)
Accuracy85.75
12
Logical ReasoningLogiQA
Accuracy74.1
11
Logical ReasoningLogiQA 1.0 (test)
Accuracy86
11
Logical ReasoningLogiQA Chinese
Pass@1 Accuracy52.4
10
Logical ReasoningLogiQA English
Pass@1 Accuracy53
10
Showing 25 of 40 rows