INDEX

Explanations

multiplicative increases

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

autoreleasepool

-0.77

 moins

-0.77

lando

-0.74

μβρίου

-0.74

万里

-0.74

 edilir

-0.73

alej

-0.73

eynman

-0.72

 bude

-0.71

万年

-0.69

POSITIVE LOGITS

 doubled

4.06

 doubling

3.73

 tripled

3.56

 doubles

3.44

 nearly

2.97

doubles

2.97

 quadru

2.97

doub

2.97

 Doubles

2.84

 double

2.81

Activations Density 0.125%