INDEX

Explanations

measurements by units

The neuron flags numeric tokens (i.e. any token containing digits).

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

 that

-1.77

 which

-1.52

 might

-1.41

 this

-1.34

 before

-1.28

and

-1.25

 where

-1.23

 also

-1.23

of

-1.22

by

-1.21

POSITIVE LOGITS

 différentes

1.33

thay

1.27

 første

1.22

 flera

1.22

 erila

1.21

 nieuwe

1.18

 lisäksi

1.18

 verschillende

1.15

 różnych

1.14

 három

1.14

Activations Density 0.023%