INDEX

Explanations

before or during

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

Éµ

-0.11

åĬ¨çĶŁæĪĲ

-0.09

iversal

-0.08

yum

-0.08

chal

-0.08

ije

-0.08

 necesita

-0.08

peat

-0.08

Winter

-0.08

alem

-0.08

POSITIVE LOGITS

 during

0.15

 beforehand

0.15

during

0.13

 prior

0.13

 During

0.12

 before

0.12

During

0.11

 durante

0.11

 BEFORE

0.10

before

0.10

Activations Density 0.853%