INDEX

Explanations

never recovered, not restarted, not back

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

 HEAD

-0.08

 vomiting

-0.08

Flo

-0.08

ibu

-0.08

Vic

-0.08

urg

-0.08

itals

-0.08

 Spread

-0.08

 Thom

-0.08

 Kirby

-0.08

POSITIVE LOGITS

 permanently

0.21

 forever

0.20

 permanent

0.17

permanent

0.15

 Forever

0.15

 perman

0.15

 irreversible

0.14

 irre

0.13

 Permanent

0.13

 never

0.12

Activations Density 0.140%