INDEX

Explanations

code statements and declarations

The neuron fires on numeric literals (both integers and decimals) in the text.

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

 your

-1.63

 with

-1.61

and

-1.44

omás

-1.42

 either

-1.36

álni

-1.35

いよいよ

-1.33

 sosteniendo

-1.31

ρύ

-1.30

我が

-1.30

POSITIVE LOGITS

geddon

1.57

hetically

1.41

⅔

1.37

athons

1.34

就连

1.33

 EXCELL

1.33

manteau

1.30

</h2>

1.30

“:

1.27

digans

1.25

Activations Density 0.030%