INDEX

Explanations

situations and introductions

This neuron fires on content-bearing nouns denoting entities, concepts, or abstract states (e.g. “value,” “situation,” “community,” “simulator,” “body,” “Muggles”).

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

and

-1.52

or

-1.48

為に

-1.42

as

-1.38

際に

-1.32

on

-1.32

づけ

-1.32

 including

-1.26

for

-1.25

 ilustracji

-1.23

POSITIVE LOGITS

<bos>

1.55

<0xA3>

1.51

がありますが

1.49

では

1.48

んじゃ

1.48

だが

1.44

ですが

1.43

そして

1.42

 Unterscheidung

1.40

があり

1.39

Activations Density 0.000%