INDEX

Explanations

was

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

 candidates

-0.87

OGND

-0.86

AddTagHelper

-0.83

 victims

-0.82

 winners

-0.82

ftagPool

-0.82

 bearers

-0.81

 sufferers

-0.80

 carriers

-0.79

rawDesc

-0.78

POSITIVE LOGITS

1.00

the

0.86

its

0.48

 ensayos

0.46

 genoux

0.44

IUrlHelper

0.42

 débats

0.40

ibly

0.40

라이

0.37

")->

0.37

Activations Density 0.001%