INDEX

Explanations

one of the

The neuron responds to multi‐digit numeric tokens (e.g. the counts like “728,” “704,” “688,” etc.).

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

itados

-1.40

inerario

-1.36

 terke

-1.34

 menggel

-1.30

 pases

-1.27

hever

-1.25

romatic

-1.25

inverno

-1.24

}*/

-1.24

 memeriksa

-1.23

POSITIVE LOGITS

is

1.92

1.69

of

1.40

 reason

1.36

 thing

1.32

 Grüßen

1.29

was

1.27

 więc

1.20

ies

1.19

1.18

Activations Density 0.013%