INDEX

Explanations

popup or pop

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

 sweep

-0.09

 resh

-0.09

 slashes

-0.08

 sweeping

-0.08

 sweeps

-0.08

 Reco

-0.08

Scr

-0.08

 dign

-0.08

 landslide

-0.08

roy

-0.08

POSITIVE LOGITS

 champagne

0.15

 cork

0.14

 Champagne

0.13

 pops

0.12

/pop

0.12

 popped

0.11

 popcorn

0.11

 unexpectedly

0.11

.Pop

0.10

.Popup

0.10

Activations Density 0.032%