INDEX

Explanations

special characters and numbers

This neuron detects percent-encoding and similar escape sequences (hex digits following “%” or “\u”) in URLs or Unicode literals.

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

-1.51

 Maybe

-1.45

to

-1.41

 zostać

-1.37

ve

-1.34

cửa

-1.33

chtend

-1.32

To

-1.31

Some

-1.31

ts

-1.28

POSITIVE LOGITS

 slika

1.55

 Aluno

1.52

 kopf

1.40

 usando

1.37

 kapture

1.34

sont

1.34

 delantero

1.32

mystic

1.30

salomon

1.29

 Produtos

1.27

Activations Density 0.001%