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

ZebraLogic

Benchmarks

Task NameDataset NameSOTA ResultTrend
Logical ReasoningZebraLogic v1.0 (test)
Cell Accuracy97.7
90
Logical ReasoningZebraLogic (test)
Grid Accuracy92.2
90
Logical ReasoningZebraLogic
Accuracy98.8
86
Logic ReasoningZebraLogic
Accuracy96
49
Logic ReasoningZebralogic
Score96.1
42
ReasoningZebraLogic
Score90.97
31
General ReasoningZebraLogic
Accuracy (%)96.3
29
Logic Puzzle SolvingZebraLogic
Accuracy60.8
25
Local Causal AttributionZebraLogic
AUPC348.71
20
Logical ReasoningZebraLogic (held-out)
Accuracy71.7
18
ReasoningZebraLogic
Avg Accuracy@144.5
12
Logic ReasoningZebraLogic
Avg Accuracy @10.817
11
Constraint Satisfaction ReasoningZebraLogic
Easy Score96.8
9
Data-to-Text GenerationZebraLogic
Schema Validity100
8
Long-horizon planningZebraLogic
Accuracy93.5
6
Logical ReasoningZebraLogic
Accuracy (Avg@8)96.3
6
Logic ReasoningZebralogic
Pass@877
6
ReasoningZebraLogic
Error Rate (%)4.01
6
Logical ReasoningZebraLogic 140 puzzles
Puzzle Accuracy130
5
General ReasoningZebraLogic
Mean@196.1
4
ReasoningZebraLogic
TPS2,245.5
3
ReasoningZebralogic
Accuracy (Zebralogic)78.7
3
Combinatorial ReasoningZebraLogic mc_mode, Rows/Cols <= 4
Average Score @6473.73
1
Showing 23 of 23 rows