INDEX

Explanations

category or topic starters

The neuron strongly activates on lines that look like section or article titles—i.e. standalone, capitalized headings.

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

 when

-1.60

 while

-1.44

Now

-1.38

 作り方

-1.38

 four

-1.38

 before

-1.35

 eight

-1.35

six

-1.30

Centre

-1.27

étiques

-1.23

POSITIVE LOGITS

 nového

1.56

 poslední

1.50

 that

1.48

 těchto

1.47

 našich

1.47



1.46

1.41

 jednej

1.36

 tohoto

1.36

 недавно

1.34

Activations Density 0.203%