INDEX

Explanations

percentage or number modifiers

The neuron is tuned to detect promotional or commercial terms and phrases (e.g. “free,” “download,” “online,” “combined with,” 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

 sexually

0.73

ି

0.72

рина

0.67

NCIA

0.62

cert

0.61

र्न

0.61

OB

0.60

cao

0.60

GUT

0.60

עה

0.60

POSITIVE LOGITS

 cambiamenti

0.84

 değiş

0.82

ارش

0.81

એસ

0.80

Shirts

0.79

ilir

0.78

 changements

0.77

 Aleksandar

0.77

 konular

0.77

 özellik

0.77

Activations Density 0.001%