INDEX

Explanations

Technical, specific writing

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 defaultstate

-0.72

 purpoſe

-0.63

 raiſ

-0.60

 Theſe

-0.60

 fhort

-0.58

 Вікі

-0.58

 Diſ

-0.57

 Beſ

-0.57

LayoutStyle

-0.57

 ſtre

-0.57

POSITIVE LOGITS

__":

0.49

 betrek

0.43

fromJson

0.43

posedge

0.43

fromCharCode

0.42

 instructed

0.41

Ecu

0.41

cault

0.41

 instructing

0.41

 isolated

0.40

Activations Density 0.004%