INDEX

Explanations

most of

This neuron selectively activates on numeric expressions—especially multi‐digit numbers and decimal values—within 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

inerario

-1.55

 одного

-1.46

ilogue

-1.38

یزات

-1.34

ichtete

-1.34

ztes

-1.30

uties

-1.29

臜

-1.29

 سایر

-1.28

᾿

-1.27

POSITIVE LOGITS

 them

1.29

ne

1.27

st

1.25

 ihnen

1.22

est

1.20

 tylko

1.20

 have

1.16

 things

1.15

 those

1.14

 slår

1.13

Activations Density 0.026%