INDEX

Explanations

numbers between dashes

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

 catalog

-0.12

/ag

-0.11

vw

-0.11

urch

-0.11

air

-0.11

 polic

-0.10

uras

-0.10

 sick

-0.10

weg

-0.10

etto

-0.10

POSITIVE LOGITS

ocr

0.08

fts

0.08

 buzz

0.08

————————————————

0.08

 kinds

0.07

ESS

0.07

 audible

0.07

uden

0.06

 Hawks

0.06

 Hamp

0.06

Activations Density 0.059%