INDEX

Explanations

code characters

This neuron activates on tokens containing numerical digits, i.e. it detects numeric values (integers or decimals) 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

it

-1.77

any

-1.32

-1.27

how

-1.25

 according

-1.23

 whether

-1.16

 should

-1.16

<bos>

-1.09

 which

-1.09

aruh

-1.09

POSITIVE LOGITS

还不错

1.26

vlá

1.19

лочка

1.19

戋

1.19

 trouva

1.18

 dentelle

1.16

izate

1.16

𓁹

1.15

σιο

1.15

可爱的

1.15

Activations Density 0.112%