INDEX

Explanations

log files, static site, donation, nuclear Iran

The neuron flags the main topical content word in each text (the key noun or technical term around which the question or discussion is centered).

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

 these

-1.74

 péché

-1.59

 навіть

-1.55

 zapato

-1.51

how

-1.48

of

-1.46

你这

-1.46

 Partagez

-1.45

 Ahnung

-1.37

年轻人

-1.35

POSITIVE LOGITS

to

1.72

tivate

1.45

}^{

1.41

*}$

1.41

}$.

1.40

uride

1.40

lapsingToolbar

1.38

ivable

1.37

}$,

1.35

you

1.34

Activations Density 0.038%