INDEX

Explanations

technical/research text

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MLLoader

-0.61

 bezeichneter

-0.57

Бахар

-0.52

ɕ

-0.51

Personendaten

-0.51

ChildScrollView

-0.50

posedge

-0.49

`;

-0.49

ivasan

-0.47

bekah

-0.46

POSITIVE LOGITS

AndEndTag

0.73

DAC

0.57

DCC

0.56

MOP

0.53

Données

0.53

 drowsiness

0.53

euse

0.52

ECE

0.52

dbl

0.52

DAC

0.51

Activations Density 0.012%