INDEX

Explanations

list separators

The neuron activates on numeric tokens and sequences—decimal numbers, section or figure numbers, and other digit‐containing strings.

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

 Japão

-1.37



-1.27

 เปิด

-1.22

 fritas

-1.22

 فقط

-1.20

 tych

-1.19

شنامه

-1.17

 unicórnio

-1.17

 图库

-1.16

ってない

-1.16

POSITIVE LOGITS

.”

1.64

can

1.48

but

1.42

its

1.41

 SizedBox

1.37

 from

1.36

↵

1.32

 during

1.30

 because

1.30

geſ

1.29

Activations Density 0.253%