INDEX

Explanations

website paths, code comments, names, and titles

The neuron responds to numeric characters (digits and decimal dots) in the text.

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

 Normalize

-0.81

畫

-0.79

 Treff

-0.73

鸽

-0.70

经营

-0.69

 Beal

-0.69

▪

-0.68

 umbi

-0.67

✅

-0.67

funktionen

-0.66

POSITIVE LOGITS

esserung

0.82

мол

0.80

vy

0.78

safety

0.78

Lus

0.77

vedať

0.77

 aseg

0.75

igation

0.75

 пациента

0.75

 misconception

0.74

Activations Density 0.006%