INDEX

Explanations

referring to paper sections and structure

This neuron detects occurrences of the word “organized” when used in phrases that describe how content or a document is structured (e.g. “organized as an …” or “organized in …”).

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

勿論

-1.63

 this

-1.59

 damaligen

-1.43

 from

-1.42

 zelfs

-1.41

 Zeitraum

-1.41

することも

-1.41

 навіть

-1.40

 günstiger

-1.38

from

-1.38

POSITIVE LOGITS

uan

1.37

ï

1.28

felf

1.27

ﺮ

1.23

are

1.23

 musik

1.22

 maand

1.21

 meeste

1.21

🌴

1.21

 solides

1.20

Activations Density 0.022%