INDEX

Explanations

through things

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 clocks

-0.11

 ourselves

-0.10

 cabo

-0.10

nek

-0.10

Pol

-0.10

 locker

-0.10

 territories

-0.10

 CLOCK

-0.10

odus

-0.09

 clock

-0.09

POSITIVE LOGITS

 veil

0.10

 spread

0.10

spread

0.09

 Spread

0.09

 coat

0.09

membership

0.09

 wiped

0.09

 membership

0.08

oles

0.08

 disgu

0.08

Activations Density 0.011%

No Known Activations