INDEX

Explanations

real challenges and possibilities

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

of

-1.55

-1.52

of

-1.51

喈

-1.49

Ｓ

-1.48

ps

-1.47

чя

-1.46

十二

-1.45

images

-1.43

了很多

-1.43

POSITIVE LOGITS

 limitada

1.86

 două

1.78

 colorida

1.77

 concentrado

1.73

 две

1.71

 marcada

1.69

dsquared

1.66

 representado

1.64

 elétrica

1.64

and

1.64

Activations Density 0.018%