INDEX

Explanations

second occurrences

The neuron detects numeric values—particularly multi‐digit numbers and decimal measurements—embedded in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

can

-1.55

for

-1.50

 from

-1.39

has

-1.35

 person

-1.31

 الأخرى

-1.22

 cannot

-1.22

 sufrimiento

-1.20

 already

-1.18

is

-1.16

POSITIVE LOGITS

 Katalog

1.41

roskop

1.38

>−

1.30

乂

1.30

ників

1.30

 aktor

1.29

 mwaka

1.27

日閲覧

1.25

,‎

1.22

 ſame

1.22

Activations Density 0.097%