INDEX

Explanations

links and foreign language words

The neuron fires on tokens that represent numeric values—particularly numbers (including floats) or digits attached to words (e.g. “41.75,” “Tuesday22,” “38.25”).

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

了一口

-1.45

it

-1.42

۱۴

-1.29

nosaurus

-1.19

 conditions

-1.17

ANDOM

-1.15

ору

-1.14

//'

-1.14

[_

-1.13

iar

-1.13

POSITIVE LOGITS

 have

1.66

包括

1.54

 для

1.52

for

1.52

 incluye

1.42

why

1.40

 algunos

1.38

 Почему

1.34

Have

1.34

 neuen

1.34

Activations Density 0.005%