INDEX

Explanations

capital improvement projects

The neuron activates on words referring to “capital” and associated large‐scale investment or project terms (e.g. capital, construction, improvement, financing, expenditures).

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

说出

-1.55

 Gables

-1.53

whose

-1.51

activités

-1.45

<h2>

-1.44

为什么会

-1.41

llllll

-1.40

Эзоҳ

-1.40

pherds

-1.37

ówczas

-1.37

POSITIVE LOGITS

and

1.84

年は

1.59

1.47

1.34

 when

1.30

歲

1.29

harap

1.28

 quería

1.23

 jeste

1.20

mları

1.20

Activations Density 0.014%