INDEX

Explanations

According to

The neuron fires on auxiliary and modal verbs (e.g. must, should, are, being, has) that signal obligation or ongoing states.

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

we

-1.22

 キャンプ

-1.22

蹕

-1.19

亖

-1.18

BoxShadow

-1.17

Menurutnya

-1.15

prende

-1.15

getWriter

-1.13

 ワゴン

-1.13

 agricultura

-1.13

POSITIVE LOGITS

1.75

 "...

1.52

 only

1.44

 contemporary

1.43

"'

1.38

:"

1.36

んだから

1.35

this

1.34

 cette

1.31

ONOM

1.30

Activations Density 0.116%