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

DiffReact

Benchmarks

Task NameDataset NameSOTA ResultTrend
Context window selectionDiffReact PDEBench (full-oracle case-level)
Lsel5
24
Regret assessmentDiffReact full-oracle case-level PDEBench
Regret Knee-0.8
16
Trajectory Rollout PredictionDiffReact f2 dense (OOD)
OOD Mean Error0.086
5
Showing 3 of 3 rows