INDEX

Explanations

private, privacy, privacy policies

This neuron specifically activates on the word “private” (especially when it forms compounds like “private label,” “private charter,” etc.).

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

 красоты

0.65

Convolution

0.64

 shoving

0.61

డ్డు

0.61

CameraOpened

0.59

Graphical

0.58

 Caval

0.57

 সিভি

0.57

 குதிக

0.57

 preempt

0.57

POSITIVE LOGITS

 priv

1.41

 Private

1.41

priv

1.35

 Priv

1.34

 Privat

1.34

 privé

1.29

 PRIV

1.24

 privat

1.21

 privado

1.20

 प्राइवेट

1.16

Activations Density 0.093%