INDEX

Explanations

qualities and outcomes

The neuron detects words that call out notable attributes, benefits, or effects—especially positive descriptors marking what’s advantageous or impactful.

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

we

-1.68

it

-1.50

 These

-1.43

 these

-1.38

 everything

-1.29

 theſe

-1.26

 Either

-1.26

 that

-1.26

who

-1.25

甚至是

-1.25

POSITIVE LOGITS

ଶ

1.39

 российской

1.37

 života

1.37

有两个

1.32

 ویکی‌پدیا

1.29

 aidé

1.29

“

1.28

 appui

1.27

 участников

1.25

𝒉

1.24

Activations Density 0.128%