INDEX

Explanations

with the

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

oft

1.60

 Belt

1.57

 Often

1.51

 Toward

1.49

 Many

1.48

 häufig

1.43

 Young

1.42

 often

1.37

 hinter

1.37

 toward

1.35

POSITIVE LOGITS

 justa

1.91

 нулю

1.77

ไม่ต้อง

1.76

껑

1.72

 '');

1.68

equality

1.68

 cero

1.68

 pongo

1.66

 whichever

1.66

 giusta

1.64

Activations Density 0.123%

No Known Activations