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

MJ-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Pairwise preference predictionMJ-Bench v1 (test)
Alignment Score69.8
7
Safety JudgeMJ-Bench
Macro-F182.23
3
Image-text AlignmentMJ-Bench
Macro F160.59
3
Showing 3 of 3 rows