INDEX

Explanations

The neuron strongly activates on tokens that are numeric literals (digits or numbers).

code definitions and keywords

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

 phénom

-1.45

not

-1.35

часы

-1.30

 menak

-1.29

 Públicas

-1.27

 zodia

-1.27

dekat

-1.21

 aprecia

-1.21

 esquí

-1.21

 emblema

-1.20

POSITIVE LOGITS

 dessa

1.43

 These

1.30

囿

1.29

 Эти

1.23

1.19

*.

1.18

 will

1.17

1.16

\\

1.16



1.16

Activations Density 0.033%