INDEX

Explanations

unsettling or disturbing

The neuron detects strongly negative or unsettling descriptive words (e.g. “disturbing,” “unsettling,” “fucked up,” “creepy”).

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

ate

-1.28

 diventa

-1.25

 conferma

-1.21

um

-1.20

 vuole

-1.17

 типу

-1.17

-1.16

 propone

-1.16

 there

-1.16

</h1>

-1.16

POSITIVE LOGITS

 mies

1.41

blende

1.33

地区的

1.30

jadikan

1.28

 ‘‘

1.27

vät

1.26

 hudeb

1.22

𖥦

1.21

 cuero

1.21

ůl

1.20

Activations Density 0.038%