INDEX

Explanations

lists of categories

The neuron selectively activates on plural nouns in list‐ or section‐style contexts (e.g. “cancers,” “activities,” “trainees”).

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

it

-1.53

but

-1.46

átku

-1.34

ógica

-1.27

(",");

-1.24

价比

-1.20

ڼ

-1.20

 only

-1.20

レッジ

-1.18

úcia

-1.17

POSITIVE LOGITS

of

1.96

1.49

 também

1.44

 también

1.25

1.23

1.22

 также

1.17

 olika

1.17

various

1.16

 всех

1.16

Activations Density 0.181%