INDEX

Explanations

concurrent adaptation, follow-up, subsequent

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

'");

-1.23

 from

-1.21

」,

-1.11

 poche

-1.07

 cannot

-1.07

 contributes

-1.03

 möj

-1.03

↵↵

-1.03

 eliminates

-1.02

ัต

-1.02

POSITIVE LOGITS

が必要です

1.36

薗

1.32

zerű

1.27

1.24

ਜ਼

1.23

ardez

1.21

 prévue

1.20

haviours

1.19

幸好

1.19

1.16

Activations Density 0.150%