INDEX

Explanations

originally planned, but changed

The neuron activates on numeric tokens with decimal points, i.e. floating-point numbers in the text.

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

the

-1.79

 различ

-1.60

曖昧さ回避

-1.38

../

-1.31

 várias

-1.30

hahah

-1.30

illés

-1.30

∷

-1.29

휠

-1.28

is

-1.26

POSITIVE LOGITS

But

1.73

but

1.72

However

1.68

 However

1.50

ーめん

1.47

 esferas

1.42

 histoire

1.42

 nhưng

1.39

 लेकिन

1.39

was

1.37

Activations Density 0.018%