INDEX

Explanations

extensive

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

of

-0.56

SourceChecksum

-0.55

 démocr

-0.51

TagMode

-0.51

 fédé

-0.49

 merve

-0.47

 cár

-0.45

 énergé

-0.45

of

-0.44

 scolaires

-0.44

POSITIVE LOGITS

\{\\

0.81

 itſelf

0.72

BibitemShut

0.72

 contextLoads

0.71

ableness

0.69

 myſelf

0.68

 fubject

0.63

 Aristi

0.62

(\<

0.61

bibfield

0.61

Activations Density 0.030%