INDEX

Explanations

list items and titles

The neuron strongly activates on tokens that begin with an uppercase letter.

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

 เขา

-1.38

 พ่อ

-1.34

deutschen

-1.24

 สอง

-1.23

 เด็ก

-1.18

になったら

-1.17

ˮ

-1.13

 แท้

-1.11

に入れる

-1.09

 douze

-1.07

POSITIVE LOGITS

 which

1.68

そして

1.37

 Because

1.28

And

1.24

 Even

1.20

ただし

1.19

1.17

It

1.16

و

1.14

があり

1.13

Activations Density 0.000%