INDEX

Explanations

flying, flight, sky, bird

The neuron selectively activates on occurrences of the word “fly” (and its capitalized or inflected forms, e.g. “Flying,” “flying”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Cheers

0.63

 incorrectly

0.62

 erroneously

0.61

 গম্ভীর

0.60

stom

0.58

 Riverside

0.55

 Département

0.55

cD

0.55

st

0.55

 Receipt

0.54

POSITIVE LOGITS

 flying

1.40

Flying

1.38

飞

1.32

FLY

1.27

flying

1.26

 Flying

1.25

飛

1.24

fly

1.23

FLY

1.19

 terbang

1.19

Activations Density 0.190%