INDEX

Explanations

encoded strings and calculations

This neuron specializes in spotting characters from Base64‐style encoded blocks (letters, digits, ‘+’, ‘/’) in things like PEM key or other binary-to-text sections.

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

殍

-1.45

盞

-1.43

¨

-1.38

樁

-1.37

我们需要

-1.34

dasarkan

-1.31

}~

-1.30

 intrusion

-1.29

后退

-1.28

 هنگام

-1.27

POSITIVE LOGITS

of

1.91

1.83

 that

1.45

ંગ

1.43

in

1.38

for

1.37

 with

1.37

 pemuda

1.35

 dunia

1.31

 også

1.30

Activations Density 0.008%