INDEX

Explanations

The neuron strongly activates on coarse profanity—especially “shit” and similar expletive intensifiers (and related words like “crap”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 breathing

0.61

 loses

0.61

 lose

0.61

 illumin

0.59

 hardwood

0.59

 incarnation

0.59

 হাফ

0.59

 zipped

0.58

 painstakingly

0.57

 intricately

0.57

POSITIVE LOGITS

Mass

0.61

cam

0.57

Priv

0.57

Tec

0.57

Scheme

0.55

List

0.55

AA

0.55

Bear

0.55

Cam

0.54

Références

0.54

Activations Density 0.010%