INDEX

Explanations

face sides and surfaces

The neuron activates on words naming object faces or surface parts (e.g. “face,” “board,” “side,” “panel”).

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

Dengan

-1.30

 alleviate

-1.29

جري

-1.27

 emphasize

-1.24

 promote

-1.24

員

-1.23

كري

-1.22

 lengthy

-1.21

أنا

-1.21

 with

-1.20

POSITIVE LOGITS

of

1.82

 faces

1.48

'*

1.45

 และ

1.41

'-

1.41

''

1.40

"'

1.40

↵

1.40

``

1.39

 contrasting

1.32

Activations Density 0.068%