INDEX

Explanations

as an example of

The neuron specializes in spotting numeric expressions—such as measurements, quantities, indexes, and decimal values—in technical 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

who

-1.59

 where

-1.55

where

-1.24

où

-1.23

 erforder

-1.16

 could

-1.14

 same

-1.13

 ähnlichen

-1.08

 advises

-1.07

を作る

-1.06

POSITIVE LOGITS

tahu

1.31

новой

1.28

aktivi

1.27

wirk

1.16

kerja

1.16

 ולה

1.15

CRY

1.14

leicht

1.14

 geda

1.13

"",

1.13

Activations Density 0.068%