INDEX

Explanations

that

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

Personensuche

-1.02

ConstraintMaker

-0.93

:✨

-0.87



-0.82

principalColumn

-0.82

SourceChecksum

-0.78

 CreateTagHelper

-0.77

parsedMessage

-0.77

ngdoc

-0.77

 AssemblyProduct

-0.76

POSITIVE LOGITS

is

1.09

’

1.02

0.99

 goes

0.94

 comes

0.76

 occurs

0.72

 happens

0.68

has

0.67

 remains

0.66

 survives

0.66

Activations Density 0.003%