INDEX

Explanations

device identifiers and attributes

This neuron flags tokens that consist of numeric characters (i.e. digit sequences).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 decken

-0.77

 Spaß

-0.77

 majesty

-0.70

 tease

-0.69

iniest

-0.67

integration

-0.65

跡

-0.65

markdown

-0.64

 Cecil

-0.64

cyto

-0.64

POSITIVE LOGITS

Chrome

0.91

 Chrome

0.90

 Topics

0.89

 Firefox

0.85

Signals

0.82

Apple

0.81

ihe

0.81

 skupiny

0.81

ENTRE

0.79

FieldDescriptor

0.78

Activations Density 0.019%