INDEX

Explanations

sentence end punctuation

The neuron selectively activates on precise numeric tokens containing decimal points (floating-point numbers).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

<0x83>

0.44

<0x86>

0.43

<0x9F>

0.40

 lazy

0.40

 IOException

0.39

indazol

0.39

 prima

0.38

IOException

0.38

 jangan

0.38

cii

0.37

POSITIVE LOGITS

 Мы

0.61

드의

0.60

of

0.57

and

0.56

 When

0.56

्स

0.54

to

0.53

affiliated

0.53

 ofthe

0.52

and

0.52

Activations Density 0.021%