INDEX

Explanations

rates of change

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ITÉ

0.59

媲

0.58

 Jagger

0.56

 düşük

0.56

 entirely

0.55

anko

0.55

 व्यावसायिक

0.55

十足

0.55

 फर्क

0.54

 рођено

0.54

POSITIVE LOGITS

 linearly

1.26

 quadratic

1.21

 monotonically

1.14

 linear

1.13

 quadrat

1.11

 monotonic

1.10

 Quadratic

1.09

 inversely

1.07

 monoton

1.02

 inverse

1.02

Activations Density 0.023%