INDEX

Explanations

Virgin, Fool, Dummies, SXSW

This neuron detects mentions of well-known brand or company names (e.g., “Virgin,” “The Motley Fool”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 errichtet

-1.30

bė

-1.25

knit

-1.24

acen

-1.24

or

-1.23

rieved

-1.23

 hâm

-1.20

automne

-1.19

culature

-1.18

 Weeknd

-1.17

POSITIVE LOGITS

—

1.34

was

1.32

 schicken

1.30

Of

1.25

dateOfBirth

1.21

 würden

1.17

—

1.16

1.13

 übri

1.13

djang

1.12

Activations Density 0.065%