INDEX

Explanations

disciples or code

The neuron fires on common English derivational suffixes (e.g. –ing, –ship, –ted) at the ends of words.

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

how

-1.71

 devrez

-1.51

 cerâmica

-1.46

but

-1.39

ほうがいい

-1.35

甞

-1.35

ünüz

-1.34

mpagne

-1.33

 outros

-1.32

でしたが

-1.32

POSITIVE LOGITS

1.40

不管是

1.37

无论是

1.37

」「

1.36

if

1.36

1.34

多

1.30

any

1.29

 schlimm

1.26

なので

1.22

Activations Density 0.109%