INDEX

Explanations

explicit content or technical settings

The neuron flags unusual, content-heavy tokens—especially proper names, numbers/ordinals, and other rare or domain-specific words.

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

enquête

0.94

 выяс

0.80

щё

0.77

новні

0.76

ের

0.74

 инфек

0.73

 Такие

0.73

ões

0.73

 Palash

0.72

 экра

0.72

POSITIVE LOGITS

سين

0.86

حن

0.82

सी

0.78

يت

0.77

سو

0.76

بة

0.75

შე

0.74

สินค้า

0.74

س

0.74

शीर

0.73

Activations Density 0.000%