INDEX

Explanations

applicable

The neuron fires on occurrences of the word “applicable.”

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

 avec

-1.37

iyi

-1.36

 buena

-1.33

);

-1.33

 kein

-1.32

is

-1.32

 orthopedic

-1.32

 símbolos

-1.31

 rendu

-1.30

 fallu

-1.30

POSITIVE LOGITS

1.62

1.58

1.53

For

1.48

you

1.46

 ataupun

1.46

1.35

 pulsante

1.34

Even

1.31

and

1.30

Activations Density 0.056%