INDEX

Explanations

campaign covering, chains added, selector switch

The neuron primarily detects numeric expressions (digits, measurements, or quantities, including decimals).

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

 about

-2.08

 those

-2.00

 which

-1.90

by

-1.89

 because

-1.81

 through

-1.70

 from

-1.69

his

-1.68

in

-1.64

 when

-1.63

POSITIVE LOGITS

 pharmacie

1.58

 accompagner

1.50

ditionally

1.50

 mélang

1.48

anwalt

1.45

苼

1.43

 trasform

1.43

 principi

1.42

 cilind

1.41

 proteine

1.41

Activations Density 0.104%