INDEX

Explanations

following "and" in parentheses

The neuron is strongly activated by closing parenthesis tokens “)” (i.e. parenthetical boundaries).

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

很赞

-1.34

ohol

-1.32

,”

-1.30

らせる

-1.28

lograph

-1.24

 zanah

-1.23

鈀

-1.22

少なくとも

-1.20

“

-1.18

POSITIVE LOGITS

}}$

1.62

of

1.48

佧

1.44

Are

1.41

 sebagainya

1.38

was

1.31

owę

1.30

 ofthe

1.26

‌

1.25

৫

1.25

Activations Density 0.035%