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

Predicting Transformer Generalization on MNIST-Transformers 60% threshold

0.897Kendall's tau

Transformer-NFN large

0.74620.785350.82450.86365Apr 26, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
0.897
2026.04
0.897
2026.04
0.895
2026.04
0.874
2026.04
0.86
2026.04
0.846
2026.04
0.822
2026.04
0.752