INDEX

Explanations

fourth chapter latest number

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

uſe

-2.94

 kemudian

-2.53

 でも

-2.52

 سپس

-2.27

утбу

-2.14

よって

-2.11

一副

-2.08

～～～～

-2.08

她就

-2.06

правки

-2.06

POSITIVE LOGITS

3.47

2.70

2.50

 disponibilidad

2.41

 escova

2.31

nt

2.31

mu

2.28

ت

2.28

最后由

2.23

ز

2.20

Activations Density 0.001%