INDEX

Explanations

search for Google

The neuron strongly activates on occurrences of the brand name “Google.”

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

codeInput

0.56

 منف

0.55

疡

0.55

 코드

0.54

DLC

0.54

 CODE

0.53

iaxial

0.53

 mision

0.53

投稿

0.52

DLC

0.52

POSITIVE LOGITS

 Google

1.55

Google

1.52

 google

1.41

google

1.38

GOOGLE

1.28

 Goog

1.26

 گوگل

1.21

 गूगल

1.20

 goog

1.17

谷歌

1.14

Activations Density 0.293%