INDEX

Explanations

location-based organizations

The neuron strongly activates on hyphenated “–based” descriptors (especially location-based organizational modifiers).

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

mıştır

-1.73

on

-1.67

to

-1.51

out

-1.43

 this

-1.43

 different

-1.41

تاح

-1.40

錳

-1.39

 those

-1.38

 considerare

-1.37

POSITIVE LOGITS

vì

1.65

鮰

1.48

am

1.48

 Japão

1.48

 ребят

1.45

os

1.45

点点

1.43

殽

1.40

tion

1.39

鉺

1.38

Activations Density 0.060%