INDEX

Explanations

code fragments

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Personensuche

-0.96

GEBURTSDATUM

-0.82

PropertyChanging

-0.78

igshid

-0.77

ьаж

-0.77

 Normdatei

-0.76

GenerationType

-0.74

 AssemblyTitle

-0.73

 تضيفلها

-0.73

ValueStyle

-0.73

POSITIVE LOGITS

at

0.82

At

0.53

 cost

0.52

At

0.52

bed

0.52

at

0.51

ad

0.51

die

0.43

 costo

0.39

 Cost

0.39

Activations Density 0.001%