INDEX

Explanations

contact information and privacy policies

The neuron fires on words and phrases that indicate contacting or reaching out (e.g. “contacted,” “to contact us,” “email us”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

詫

-0.81

 added

-0.70

arder

-0.69

 underworld

-0.69

阻止

-0.68

 order

-0.68

 चित्र

-0.68

inama

-0.68

 alltså

-0.67

桝

-0.67

POSITIVE LOGITS

privacy

1.14

 privacy

1.09

Privacy

0.98

 Privacy

0.97

ONEY

0.94

்கு

0.94

gd

0.92

ígeno

0.91

 jät

0.90

āci

0.89

Activations Density 0.018%