INDEX

Explanations

module exports

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

 spectacular

-1.54

 hugely

-1.48

 REALLY

-1.42

 magnificent

-1.38

 joyful

-1.38

rô

-1.35

 breathtaking

-1.34



-1.34

 cinq

-1.30

 juſt

-1.30

POSITIVE LOGITS

 gives

1.73

 allows

1.70

 provides

1.64

 what

1.62

Surprisingly

1.52

 enables

1.50

 through

1.49

 believes

1.45

 seems

1.43

 demonstrates

1.43

Activations Density 0.043%