INDEX

Explanations

Down/Go

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

 CURIAM

-0.83

didSet

-0.77

 समीक्षक

-0.75

PhysRevD

-0.74

awtextra

-0.71

/\.

-0.70

 Scribd

-0.70

 AssemblyCulture

-0.69

tisserie

-0.68

 ProtoMessage

-0.68

POSITIVE LOGITS

 down

0.66

down

0.44

Down

0.43

und

0.41

 DOWN

0.38

GO

0.36

 Down

0.35

go

0.35

Go

0.35

ent

0.34

Activations Density 0.013%