INDEX

Explanations

wheels and rims

The neuron consistently responds to occurrences of the word “wheel” (in headings, proper names, and technical contexts).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

“

-1.37

’

-1.37

 chinos

-1.30

并未

-1.29

-1.28

的模样

-1.27

Its

-1.24

 Advertisement

-1.23

 “(

-1.21

的她

-1.21

POSITIVE LOGITS

 similaires

1.70

Ⴈ

1.63

ходим

1.63

 और

1.52

 και

1.51

τηση

1.50

jähriger

1.48

هایت

1.45

 jakość

1.45

明月

1.45

Activations Density 0.031%