INDEX

Explanations

list item markers or numbers

The neuron selectively activates on numeric tokens—whether list indices in parentheses, decimal values, or integer measurements—i.e. it finds numbers.

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

𝗼

-1.21

 اشار

-1.21



-1.20

ὠ

-1.19

 particolarmente

-1.18

 joyería

-1.18

subsec

-1.18

 diputados

-1.18

对你

-1.16

 schlaf

-1.13

POSITIVE LOGITS

for

1.76

🩹

1.45

we

1.44

has

1.34

in

1.31

 Even

1.31

 mildly

1.28

by

1.28

 vaksinasi

1.27

 there

1.25

Activations Density 0.005%