INDEX

Explanations

attention

attention mechanisms

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

લાઇ

0.43

ার্জ

0.39

人心

0.39

ace

0.38

άζ

0.38

}>{

0.37

ас

0.37

 تحد

0.37

मेय

0.36

AUS

0.36

POSITIVE LOGITS

 attention

1.86

attention

1.78

Attention

1.70

 Attention

1.69

 внимание

1.59

 внимания

1.54

 aandacht

1.52

 atención

1.50

 atenção

1.48

 Aufmerksamkeit

1.46

Activations Density 0.176%