INDEX

Explanations

foreign language words

The neuron selectively activates on numeric tokens—especially numbers with decimal points or fractional values (e.g. “24.5,” “13.438,” “25.25,” etc.).

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

 your

-1.77

if

-1.76

 когда

-1.66

You

-1.60

you

-1.50

to

-1.45

when

-1.45

（

-1.40

-1.34

If

-1.32

POSITIVE LOGITS

ㅊ

1.45

 cokelat

1.38

 royale

1.35

 boute

1.35

 ruines

1.34

过来

1.33

 Pyrénées

1.31

ってない

1.31

 sebagainya

1.30

 cherokee

1.28

Activations Density 0.060%