INDEX

Explanations

when you want

The neuron fires most strongly on the first content word of a new sentence or major clause—that is, sentence-initial tokens.

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

still

-1.53

 bougies

-1.43

its

-1.41

 jeszcze

-1.40

 kroner

-1.38

 ainda

-1.36

 composants

-1.34

 verres

-1.34

 permitiendo

-1.33

 arbres

-1.31

POSITIVE LOGITS

 this

1.81

now

1.56

뿐

1.44

 навіть

1.42

それでは

1.38

 зараз

1.37

 without

1.35

1.30

TextEditing

1.30

困難

1.29

Activations Density 0.049%