INDEX

Explanations

presentedrenowned nurserydelicious steamcombinesexhausted ofendless seriesrejectcartilagesreceivedmanner

The neuron consistently activates on numeric tokens (especially numbers and decimals), i.e. it detects occurrences of numeric values 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

 ngayon

-1.48

۰۱

-1.39

 cantora

-1.25

 stí

-1.25

 éve

-1.19

 aussit

-1.18

ついで

-1.18

 fraî

-1.16

 skvě

-1.16

 pagkak

-1.16

POSITIVE LOGITS

for

1.63

to

1.58

如果

1.31

२

1.27

如果你

1.26

每个

1.25

it

1.23

It

1.15

but

1.14

signol

1.14

Activations Density 0.148%