INDEX

Explanations

at all

The neuron fires on words that qualify or intensify statements—e.g. quantifiers, negations, and strong modifiers like “all,” “not,” “totally,” “important,” “working,” etc.

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

…,

-1.63

ure

-1.55

Cat

-1.53

may

-1.49

var

-1.48

do

-1.48

ha

-1.48

–

-1.45

um

-1.45

generally

-1.42

POSITIVE LOGITS

 menulis

1.50

であり

1.38

for

1.37

 membag

1.37

legung

1.36

itemName

1.35

or

1.34

 mengikut

1.34

ющая

1.30

 wobec

1.30

Activations Density 0.004%