INDEX

Explanations

references to specific types of code or configuration files commonly shown as bracketed or tokenized placeholders

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

enumi

-1.16

rungsseite

-1.13

EndInit

-1.04

+#+

-1.01

wapV

-0.97

uxxxx

-0.97

UserScript

-0.97

ArgsConstructor

-0.94

featureID

-0.94

EDEFAULT

-0.93

POSITIVE LOGITS

](

2.28

)](

1.41

"](

1.12

[](

0.62

]()

0.60

</

0.59

なく

0.58

>(

0.56

httphttps

0.52

[](

0.52

Activations Density 0.015%

No Known Activations