INDEX

Explanations

shops, towns, and businesses

The neuron strongly activates on the main nouns naming venues, organizations, or key objects (e.g. “newspaper,” “club,” “complex,” “school,” “plant,” “show,” “pot”) that serve as the central subjects of news‐style event descriptions.

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

了一会儿

-1.59

pecabe

-1.57

 estekak

-1.55

𓆝

-1.42

勫

-1.42

ープル

-1.37

 usual

-1.36

 insanlar

-1.35

ödie

-1.33

POSITIVE LOGITS

 that

2.00

 they

1.55

 called

1.51

 campion

1.44

of

1.41

 другими

1.38

 then

1.35

 them

1.31

锲

1.30

ógico

1.29

Activations Density 0.287%