INDEX

Explanations

than any other

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

 formidable

-0.10

ANY

-0.10

 types

-0.10

fel

-0.10

ANY

-0.10

İ

-0.10

cru

-0.09

umph

-0.09

 somewhat

-0.09

ses

-0.09

POSITIVE LOGITS

 previous

0.16

 previously

0.15

 imaginable

0.14

previous

0.14

 conceivable

0.12

ive

0.11

 Seen

0.11

 imagined

0.10

 mere

0.10

(previous

0.10

Activations Density 0.020%