INDEX

Explanations

Japanese patent numbers

The neuron fires on numeral tokens—i.e. sequences of digits (numbers, years, version or ID 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

ֵי

-1.46

one

-1.38

 жы

-1.28

TRIBUT

-1.28

 finestra

-1.26

 官网

-1.26

狷

-1.23

𝘜

-1.23

ruvate

-1.23

two

-1.22

POSITIVE LOGITS

 также

1.45

 pudiera

1.32

叭

1.26

 المزيد

1.24

において

1.23

沨

1.21

гают

1.19

なので

1.19

 dejó

1.17

幸好

1.16

Activations Density 0.012%