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

RAVEN

Benchmarks

Task NameDataset NameSOTA ResultTrend
Abstract Visual ReasoningRaven
Accuracy98.7
25
Abstract Visual ReasoningRAVEN v1 (test)
Average Accuracy93.6
22
Relational ReasoningRAVEN (test)
Average Accuracy92.5
5
Macro-Average (Simulation Tasks)Raven Bench
Success Rate (SR)88.57
4
Assembling of Kits (d)Raven Bench
Success Rate95.21
4
Finding Horizontal Symmetry Letters (c)Raven Bench
Success Rate80.01
4
Placement of Blocks in Bowls (b)Raven Bench
Success Rate (SR)92.01
4
Stacking of Blocks in Zone (a)Raven Bench
Success Rate (SR)87.82
4
Showing 8 of 8 rows